In Paris this week on an official visit, Azerbaijan’s autocratic President Ilham Aliyev has already scored one photo op. Anyone reading yesterday’s Azeri media could see dozens of photos of Aliyev posing with leaders of top French companies, including Airbus, Suez, and Credit Agricole.

Azerbaijan's President Ilham Aliyev (L) shakes hands with his French counterpart Francois Hollande as they visit a local French school under construction in Baku, May 11, 2014.

© 2014 Reuters

Today, President Hollande will receive President Aliyev and host an official dinner at Palais de l’Elysee. Again, Parisian photo ops abound. But amid the flashing cameras, one has to wonder where Azerbaijan’s repression of critics and the jailing of opponents fits in the new relationship between Paris and Baku?

In the past few years, Azerbaijani authorities have aggressively gone after the country’s once vibrant civil society, jailing dozens of activists, journalists, and political opponents. It also adopted draconian legislation making it virtually impossible for independent non-governmental organizations to operate.

One year ago, as Azerbaijan’s economy started to suffer from falling oil prices, several of those detained on political grounds were released. That was an important first step, but hopes for progress were short-lived.

Many of those released face travel bans or obstacles to their activities. Dozens are still locked up on political grounds, including opposition activist Ilgar Mammadov, despite repeated calls by the Strasbourg-based Council of Europe for his immediate release. And more activists have been thrown in jail. Recently, one of the country’s most popular journalists and bloggers, Mehman Huseynov, was sentenced to two years in prison for allegedly defaming the police, in response to his brave public denouncement of the police abuses he suffered.

When visiting Paris, Brussels, or other European capitals, President Aliyev hopes to get more business opportunities and investment in Azerbaijan. But he prefers to ignore that the people of Azerbaijan want human rights protections, transparency, and good governance. Those standing up for these values are routinely exposed to attacks and harassment.

Yet what more clear message that Azerbaijan’s crackdown cannot be ignored by potential investors than last week’s decision by the Extractive Industries Transparency Initiative (EITI), an international coalition promoting better governance of resource-rich countries, to suspend Azerbaijan – precisely because of its actions against civil society.

President Hollande should reject a narrative that only finance and economy matter in Azerbaijan. Human rights should be as central to France’s foreign policy as other topics.

Hollande should publicly call for the release of Ilgar Mammadov and all those detained in retaliation for their activism and criticism. A failure to explicitly support human rights principles would be the worst message to those unjustly waiting behind bars.

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am

Arvind Ganesan is the director of Human Rights Watch’s Business and Human Rights Division. He leads the organization’s work to expose human rights abuses linked to business and other economic activity, hold institutions accountable, and develop standards to prevent future abuses. This work has included research and advocacy on awide range of issues includingthe extractive industries; public and private security providers; international financial institutions; freedom of expression and information through the internet; labor rights; supply chain monitoring and due diligence regimes; corruption; sanctions; and predatory practices against the poor. Ganesan’s work has covered countries such as Angola, Azerbaijan, Burma, China, Colombia, the Democratic Republic of Congo, Equatorial Guinea, India, Indonesia, the United States, and Nigeria. His recent research has focused on predatory lending practices and governance issues on Native American reservations in the United States. He has written numerous reports, op-eds, and other articles and is widely cited by the media.

Ganesan has also worked to develop industry standards to ensure companies and other institutions respect human rights. He is a founder of the Voluntary Principles on Security and Human Rights for the oil, gas, and mining industries and is a founding member of the Global Network Initiative (GNI) for the internet and telecommunications industries, where he also serves on the board. Ganesan has helped to develop standards for international financial institutions such as the World Bank, and regularly engages governments in an effort to develop mandatory rules or strengthen existing standards such as the Kimberley Process. He serves on the board of EGJustice, a nongovernmental organization that promotes good governance in Equatorial Guinea, and is a member of the International Corporate Accountability Roundtable (ICAR)’s steering committee.

Before joining Human Rights Watch, Ganesan worked as a medical researcher. He attended the University of Oklahoma.

Posted: January 1, 1970, 12:00 am

Children pan for gold along the mercury-contaminated Bosigon River in Malaya, Camarines Norte.

© 2015 Mark Z. Saludes for Human Rights Watch

(London) A major jewelry industry group has made significant improvements to its process for certifying its members when it comes to ensuring human rights risks are addressed in supply chains, 27 civil society groups said today in an open letter to the industry group. But the organizations highlighted serious concerns with how the certification process will be carried out and urged the industry group, the Responsible Jewellery Council (RJC), to address them.

The Responsible Jewellery Council includes over 1,100 member companies along the jewelry supply chain. In April 2019, the RJC published a new version of its certification standard, the “Code of Practices,” following consultation and input from industry and civil society groups. All RJC members are required to comply with the standard and are audited against it two years after joining.

“It is good news for human rights protection that the Responsible Jewellery Council will now require member companies to put into practice the international norm on responsible sourcing,” said Juliane Kippenberg, associate child rights director at Human Rights Watch. “But it is a source of concern that the Responsible Jewellery Council has put implementation for the diamond industry on a slower track, risking coming across as soft on that part of the industry.”

The revised standard now requires Responsible Jewellery Council members to exercise due diligence over their supply chains in accordance with the leading international norm on responsible mineral sourcing, the Due Diligence Guidance for Responsible Supply Chains of Minerals by the Organization for Economic Cooperation and Development (OECD). Under the OECD guidance, companies must take the following five steps: set up management systems for due diligence; identify human rights risks in their supply chains; respond to the risks found; carry out an independent, third-party audit; and report publicly about the steps taken.

The RJC’s current phase-in plan for carrying out the new certification standard foresees that diamond and colored gemstone companies’ existing practices will not be fully assessed against the standard until April 2021 at the earliest. Certification can be granted for up to three years, meaning that some members may not be fully checked for compliance until 2024.

Furthermore, a review of the standard and its assessment process is scheduled for after April 2021, creating additional uncertainty over whether and when diamond and gemstone companies will be checked for full compliance. Companies in other jewelry supply chains only have a transition period of one year, until April 2020.

Other concerns highlighted by the groups include what they consider to be inconsistent reporting requirements under the new standard. One provision calls for public reporting, while another one provides for reporting solely to stakeholders. The groups also expressed concern that the certification process remains opaque and “could allow member companies to pursue irresponsible business practices.”

The groups recommended requiring companies to publish audit summaries, including information on all facilities visited, areas of non-compliance, a description of any identified risks, and the specific measures taken to assess and mitigate risks. They also said that the Responsible Jewellery Council should ensure that all members up for certification beginning in April 2020 are audited against the new Code of Practices, including the full five-step framework required by the OECD’s Due Diligence Guidance.

“The Responsible Jewellery Council should make its audit program stronger and more transparent,” Kippenberg said. “It is essential for consumers to understand what steps have been taken to ensure that the jewelry they buy wasn’t made at the expense of the workers’ human rights and safety.”

Posted: January 1, 1970, 12:00 am

Barkan, located in the occupied West Bank, is an Israeli residential settlement and industrial zone that houses around 120 factories that export around 80 percent of their goods abroad. In the background is the Palestinian village of Qarawat Bani Hassan.

© 2004 David Silverman

There is a bitter irony to the White House's newly released economic development plan for the Occupied Palestinian Territory. Titled "Peace to Prosperity," the plan was devised by Jared Kushner, Donald Trump's son-in-law and senior adviser, along with David Freedman, the U.S. ambassador to Israel, and Jason D. Greenblatt, a special envoy for international negotiations. It seeks to raise $50 billion, mostly from Arab countries, around half of which would be used to develop the Palestinian economy, while the rest would go to Palestinians living in Egypt, Jordan, and Lebanon.

But the plan avoids addressing key obstacles to economic development: the closure of Gaza and, in the West Bank, Israeli settlements, which are illegal under international humanitarian law, and a two-tiered discriminatory system that treats Palestinians and the settlers separately and unequally. The lack of economic growth is not just a byproduct of these abuses, but the result of deliberate Israeli policies.

When Israel occupied the West Bank and Gaza in 1967, it cut the territories off from their previous trading partners. In 1968, Moshe Dayan, then the Israeli defense minister, said "[W]e can create economic integration... We should connect the two [Palestinian and Israeli] entities, if we, on our part and for ourselves, do not want to sever connections with these areas." But integration did not connote equality. On the contrary, Israel has continued to expand settlements and further entrench its discriminatory system against Palestinians, even when Israel partly reversed integration after 1994, following the Oslo Accords.

Over the past five decades, Israel has used its control over Palestine's borders, land, and water to build lush residential communities for more than 600,000 Israeli settlers and 19 industrial zones, in violation of the laws of occupation, while severely limiting Palestinians' access to their own natural resources and the permits needed to develop them. In 1987, Ariel Sharon, who was then the industry and trade minister, told the Knesset that his policy is to "strictly examine" requests by Palestinians to build factories, and "comprehensively take into account Israeli industries, the needs of the Israeli market, and the potential for export." He added that the threat of Palestinian competition "mandates the establishment of [Israeli settlement] industry."

As a practical matter, this has meant, for example, that the Israeli government grants its citizens and foreigners permits to build factories in the West Bank on land it has unlawfully seized – often awarding generous subsidies to encourage investment – while systematically denying such permits to Palestinians, even for land they own. This inverts Israel's international law obligations, discriminating against the people for whose benefit the occupying country is required to administer the territories and privileging those whom the laws of occupation prohibit from living there in the first place.

The case of West Bank stone quarries illustrates how Israel's discriminatory restrictions cost the Palestinian economy $241 million annually, according to the World Bank. Israel licenses 11 settlement-operated quarries in the West Bank, which supply around one quarter of its gravel market, despite this exploitation of resources in occupied territory violating international humanitarian law. One of these quarries is owned by Hanson, a subsidiary of Germany-based Heidelberg Cement. Israel's Civil Administration granted the Heidelberg subsidiary a permit to quarry on land that it seized from the Palestinian village of Zawiyeh.

The ease with which these settlement quarries operate contrasts with Israel's virtual ban on issuing Palestinian permits for quarries for the last three decades. Israeli authorities, for example, stopped renewing permits for quarries around Beit Fajar, a town of about 13,500, 10 kilometers south of Bethlehem. In 2010, 80 percent of the town's jobs were in the stone industry spread among 150 stone workshops and 40 quarries. But in recent years, the authorities stopped renewing permits for the few quarries they had allowed to continue operating. Quarry owners who continue to operate often face hefty fines and the confiscation of expensive equipment, in addition to difficulties transporting their product due to delays at the hundreds of checkpoints and road obstacles scattered across the West Bank.

Many Palestinian industries have a similar story. Israeli policies stunt their development, while helping unlawful settlement industries to thrive. According to the World Bank, Israeli restrictions in Area C of the West Bank, the area under exclusive Israeli security control, cost the Palestinian economy $3.4 billion per year.

If the White House wants to bring peace through prosperity, it should press Israel to end its unlawful and discriminatory policies that are helping to strangle the Palestinian economy.

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am

Bangladeshis work at Snowtex garment factory in Dhamrai, near Dhaka, Bangladesh, April 19, 2018. 

© 2019 AP Photo/A.M. Ahad

The German Ministry of Economic Affairs and Energy is trying to weaken measures that would track how well companies in the country identify and respond to possible human rights abuses in their supply chains. The coalition government should stand firm at a meeting of secretaries of state taking place today and adopt a monitoring system that holds German companies to rigorously high standards when it comes to sourcing materials responsibly. The companies should ensure that their supply chains are free of human rights abuses from start to finish – in line with internationally recognized norms.

“The Economics Ministry is putting forward a proposal that would make it far too easy for companies to be categorized as complying with international human rights standards when they are not doing the job,” said Juliane Kippenberg, associate children’s rights director at Human Rights Watch. “There is a risk that the monitoring system will be misused as a political tool to avoid tougher government measures on companies – most importantly, the passage of a much-needed law on supply chains.” A supply chain is composed of all steps needed to make a product – from obtaining raw materials to shipping, manufacturing, and retail.

The parties making up Germany’s coalition government have agreed between themselves that if the country’s large companies don’t voluntarily address supply chain and other human rights abuses sufficiently by 2020, the government will consider creating laws that require them to do so. Companies would have to identify, mitigate, and account for the human rights impact of their activities.

Whether the government moves forward with proposing a law on supply chains depends therefore, in large part, on how thoroughly the government is monitoring companies’ performances. However, the proposed monitoring system has already been derailed by disagreements among government officials. A questionnaire for self-reporting on the topic was supposed to have been sent out to companies based in Germany with more than 500 employees for them to respond to between May and July. But this process has stalled, and the questionnaires have yet to be distributed because the Economic Affairs Ministry disagrees with the plan.

The ministry is proposing a monitoring system that would allow the government to categorize more companies as complying with government standards for responsible sourcing. It suggests that rather than only two categories – “compliant” and “non-compliant” –, there should be four, including those “with compliance plans” and “partially compliant.”

Germany’s 2016 National Action Plan on Business and Human Rights foresees a “robust” monitoring system to assess company performance, which is far from the Economic Ministry’s current proposal.

According to the International Labour Organization (ILO), more than 450 million people work in supply chain-related jobs. Human Rights Watch has documented labor rights abuses such as child labor, forced labor, hazardous working conditions, attacks on trade unionists, and other serious human rights abuses in global supply chains.

For example, child laborers and adults in Ghana, the Philippines, and many other countries risk ill-health and death when mining gold in unstable pits and processing ore with toxic mercury. Human Rights Watch has also documented labor abuses in global garment supply chains, including excessive or forced overtime work, denial of breaks, pregnancy discrimination, and attacks on trade unionists.

While a number of German apparel companies have joined the German Partnership for Sustainable Textiles, many of these companies still do not adopt basic human rights due diligence good practices, such as supply chain transparency. In addition, changes in apparel companies’ purchasing practices and the introduction of quality grievance redress measures for workers in their global supplier factories are critical to mitigate labor abuses in global supply chains. Human Rights Watch has also documented that German companies such as the jeweller Christ and the clothing brand KiK, do not have human rights safeguards in their supply chains.

“The Economics Ministry’s delaying tactics and whitewash proposals are a disgrace,” Kippenberg said. “The government should show that respect for international human rights norms at home and abroad and Germany’s economic interests are not mutually exclusive, and that support for good practices can enhance economic growth.”

Posted: January 1, 1970, 12:00 am

Resident walks through the streets of Macomia on June 11, 2018 in Macomia, Northern Mozambique. Cabo Delgado, a northern province expected to become the center of a natural gas industry after several promising discoveries, has seen a string of assaults on security forces and civilians since October 2017.

© 2018 Emidio Josine/AFP/Getty Images

This week, Mozambique’s capital, Maputo, received more than ten African heads of state and over a thousand US government representatives and entrepreneurs for the US-Africa Business Summit. The summit is a platform for exploring business opportunities and advocacy for US-Africa trade and investment policies. On Tuesday, just before the summit, the US energy firm Anadarko announced the construction of Mozambique’s first onshore gas project.

While the meeting is important for promoting investment in the country, its agenda included no discussions about the links between business, insecurity, and human rights in the areas being considered by investors. For example, the littoral of Cabo Delgado province, where Anadarko will explore one of the largest gas reserves in the world, has endured attacks from a suspected armed Islamist group since October 2017. Those attacks and the government’s response to them have led to massive human rights abuses including killings, enforced disappearances, arbitrary arrests, and burning of houses. The authorities have also barred the media from the region, heightening uncertainty about the situation.

Mozambican civil society groups play a crucial role in monitoring corporate social responsibilities, helping communities benefit from investments and reducing tensions between communities and businesses. The US-Africa Summit should be used as a platform to address those issues and focus on initiatives to ensure that human rights in areas of investment are protected.

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am

(Paris) – Equatorial Guinea is seeking to rejoin a group that works to curb corruption in resource-rich countries, Human Rights Watch, EG Justice, and Publish What You Pay said today. The Extractive Industries Transparency Initiative (EITI) requires transparency around oil, gas, and mining revenue and activities and respect for civil society. The group delisted Equatorial Guinea in 2010, in part over concerns about its repression of civil society.

Before accepting an application, the EITI board should carefully scrutinize whether it includes concrete steps to address the government’s long history of civil society repression and to create an enabling environment for citizens to engage with resource governance, the groups said. The International Monetary Fund should similarly scrutinize the application before deciding whether it satisfies a requirement that would make Equatorial Guinea eligible for a loan.

“If Equatorial Guinea follows EITI’s requirements, it would bring much-needed transparency to its oil sector and protect people’s right to engage in public debate, including about government mismanagement and corruption,” said Sarah Saadoun, a business and human rights researcher at Human Rights Watch. “But the Equatorial Guinean government’s unrelenting attacks on critics calls into question whether this is a good-faith effort to apply to EITI or a check-the-box exercise to qualify for an IMF loan.”

Equatorial Guinea is a small, oil-rich Central African nation led by President Teodoro Obiang Nguema Mbasogo, who has amassed enormous wealth for himself and his family during his 40 years in power. Gabriel Mbaga Obiang Lima, the president’s son who is also minister of mines and hydrocarbons, delivered a speech at EITI’s Global Conference in Paris on June 18, 2019, during which he announced his government’s intention to apply to rejoin the initiative. If accepted, it would fulfill the final requirement to make the government eligible for a loan package from the IMF under one of its current programs.

Any IMF loan without genuine and sweeping governance reform raises significant concerns given the government’s history of corruption and gross financial mismanagement, the groups said. These conditions directly contribute to the country’s significantly poorer health and education outcomes than regional averages, even though it has the highest per capita income on the African continent.

To apply to EITI, a country is required to establish a national committee that includes government officials and business and civil society representatives to develop a plan to comply with the initiative’s standards. The plan must disclose key information related to governance of oil, gas, and mining enterprises and confirm that the government fosters “an enabling environment for civil society.” Equatorial Guinea has established such a committee, though its civil society members risk reprisals if they speak against the government.

Equatorial Guinea joined the EITI in 2007 but was delisted by the EITI board in 2010. The government’s well-documented harassment of a good governance advocate, Alfredo Okenve, is emblematic of the government’s intolerance of dissent. Okenve represented civil society on the EITI steering committee from 2007 to 2010 and again between 2015 and 2017 and is the vice president of one of the country’s few independent human rights organizations.

In May 2010, he was removed from his position at Equatorial Guinea’s National University after critiquing the government’s transparency record at an event in Washington, DC. He resigned from EITI in 2017, after police held him in a police station for two weeks until he agreed to pay an arbitrarily imposed fine of more than US$3,000. In October 2018, four men, who appear to have been undercover security officials, dragged him from his car at gunpoint, drove him to a remote location, and beat him severely. Three months ago, the police prevented him from leaving the country.

The government’s reprisals against Okenve garnered international attention, but there are many similar instances of repression against other civil society members that have largely been unreported. For example, a good governance activist who is also a member of a political opposition group, Joaquin Elo Ayeto, has been in preventative detention at the Black Beach prison in Malabo since February 25, apparently for making comments critical of the government’s spending priorities.

Two civil society representatives on the national EITI committee, who asked not to be quoted by name for fear of reprisals, say that civil society repression has grown worse in recent years. One said that: “creating an ‘enabling environment’ for civil society is not about us in the EITI committee. It’s about civil society more broadly having the space to operate.” The other said: “The government doesn’t respect civil society and doesn’t engage with them beyond EITI. No progress is being made. They’re just trying to comply with EITI for the sake of the [IMF] loan.”

“EITI’s legitimacy as an anti-corruption tool is diminished if governments can abuse the process to bolster their image without the requisite political will or a genuine commitment to respect civil society,” said Tutu Alicante, executive director of EG Justice, an organization promoting human rights and good governance in Equatorial Guinea. “It would be irresponsible to give Equatorial Guinea the benefit of the doubt when an anti-corruption activist has been wallowing in a crowded prison cell since February, apparently for daring to criticize government spending.”

Posted: January 1, 1970, 12:00 am

Kathy Kraninger, director of the Consumer Financial Protection Bureau, takes questions from the House Financial Services Committee's biannual review of the CFPB, on Capitol Hill in Washington, Thursday, March 7, 2019.

© 2019 J. Scott Applewhite/AP Photo


The US Consumer Financial Protection Bureau’s (CFPB) announcement last Thursday that it will delay implementing key parts of its rule regulating payday, vehicle title, and other high-cost installment loans by 15 months, will leave borrowers vulnerable to serious threats posed by these loans, which often carry triple-digit interest rates.

The rule, finalized in November 2017 after extensive consultations with many stakeholders, was meant to better protect vulnerable consumers from payday and other small-dollar, short-term predatory loans. But Kathy Kraninger, the recently appointed director of the CFPB, proposed major changes to the payday rule in February this year, just two months after her appointment. In addition to proposing a delay to the rule’s implementation, the CFPB also proposed repealing key borrower protections, including a requirement that lenders ensure that borrowers can actually repay their loans.

Borrowers, who often live paycheck to paycheck, can fall into debt traps when they are unable to keep up with payments. They may have to take out new loans to repay the old ones, creating a cycle of growing debt. In some cases, people can end up owing much more than they originally borrowed, and may forego food, rent, or other basic needs to pay back their loans.

Human Rights Watch joined 130 consumer, civil and human rights, labor, and community organizations in opposing the CFPB’s proposed delay in implementing the rule. Human Rights Watch also opposed the rollback of critical consumer protections, highlighting our research into the impacts of predatory, high-interest loans on Native Americans.

The CFPB should require companies to comply with the payday rule, as originally intended, by August 2019. The CFPB should also retain all parts of its rule, including the ability-to-pay requirement on lenders, and stop putting vulnerable borrowers at the mercy of loans they will struggle to repay.

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am

Campaigners outside the Royal Courts of Justice in London supporting the legal challenge against the Government's Universal Credit welfare scheme. 

© 2019 Stefan Rousseau/Press Association via AP Images

The UK government has embarked on an ambitious strategy to digitise and automate the delivery of welfare services, but misplaced reliance on these technologies is threatening to further unravel the social safety net.

Last month, the United Nations expert on poverty and human rights and Human Rights Watch published their respective findings on the state of poverty in the UK, documenting the toll of austerity policies on the rights of the poorest people in British society. The UN report, written by Professor Philip Alston, warned in particular that the “British welfare state is gradually disappearing behind a webpage and an algorithm.”

The country’s transition to digital welfare is closely linked to its austerity-motivated efforts to slash spending on welfare and other essential public services, while at the same time fundamentally restructuring the social security system. In 2012, the government created Universal Credit, which consolidated six social security benefits into a single benefit paid monthly.

Universal Credit is designed to be “digital-by-default”: most beneficiaries are expected to apply for and manage their benefits through an online portal, and offline means of obtaining benefits have been pared back. The Department of Work and Pensions, which oversees Universal Credit, has claimed that the online delivery of benefits would make them “easier to access” and “cheaper to provide.”

In practice, however, the rollout of Universal Credit reveals a mismatch between the government’s digital aspirations and the rights of some of the country’s most vulnerable people. The UN report found that many claimants who lack digital literacy skills or cannot afford internet access at home experience difficulties claiming their benefits online. The burgeoning demand for digital assistance with Universal Credit has transformed public libraries, which provide free computer services, into makeshift crisis centres – a role they are ill-equipped to undertake in light of staff shortages and their own budget cuts.

The government’s official oversight and audit body has raised serious questions about whether this system in fact saves money or costs more, and whether the government has provided adequate support to local authorities helping people obtain their benefits. This isn’t news to the government: parliamentarians and local government representatives have consistently flagged that the resources to support a “digital by default” system are not adequate.

The drive to automate critical aspects of Universal Credit has heightened the risk that people might be wrongly denied essential benefits. As of March 2018, only 38% of welfare claimants had been successfully verified under the government’s initiative to automate identity checks through services provided by the private sector. These services depend on information such as financial and phone records and immigration documents. But public assistance groups have found that many claimants lack such documentation or struggle to provide it online.

Errors in the automated processing of benefits have also sowed confusion and financial hardship among claimants. Late or incorrect payroll information provided by their employers has caused delays and reductions in benefit payments, which can take weeks to rectify. These errors are potentially catastrophic for claimants already struggling to feed their children and make ends meet. Although the government has proposed policy changes to fix these errors, it has admitted that it does not know “how this will work in practice.”

The UK is not alone in its struggles with digital welfare. A growing number of countries have botched similar efforts. In India, millions of poor families have been denied food rations because they were unable to satisfy digital identification requirements. When municipal governments in the United States and Canada replaced caseworkers with automated systems to process welfare claims, the underpayment and denial of benefits surged. Thousands of welfare claimants were wrongly accused of benefits fraud in Michigan because of design flaws in the state’s automated fraud detection system.

In the UK, these failures are increasing the strain on welfare claimants, who are already suffering from sanctions for falling short of work requirements and from reductions in their benefits over the last decade.

Despite these red flags, the government is pressing on with its digital transformation of the welfare system. A “fully automated” system for detecting fraud is also in the works. If the government does not first resolve ongoing failures with the technologies it has already rolled out, it risks repeating them. These upgrades should also not distract from the urgent task of re-evaluating the underlying premises of Universal Credit, and the slew of benefit caps, freezes and penalties enacted in its name.

Without welfare reform that addresses the root causes of poverty, hunger and homelessness, automation and other technologies may only serve to exacerbate their effects.

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am

As leading labor, human rights, and environmental organizations and members of the Thai Seafood Working Group, we are deeply concerned that the Marine Stewardship Council’s new Chain of Custody Certification will not be effective in identifying, preventing, or protecting seafood workers from labor rights violations. Moreover, the new requirements for assessing risk and preventing child labor and forced labor will not provide buyers and retailers with the assurance that child labor and forced labor are not present in their supply chains.

In March 2019, the Marine Stewardship Council (MSC) released new requirements for its Chain of Custody Certification (CoC) for on-shore seafood operators, which include provisions to address forced and child labor in seafood supply chains, for the first time.[1]

The Thai Seafood Working Group and other concerned stakeholders engaged proactively in the public consultation process to ensure that the labor requirements for the CoC Certification were effective and robust. These stakeholders thoughtfully made recommendations for improving MSC’s labor requirements. MSC incorporated very few of these suggestions.

MSC’s labor requirements are inefficient, ineffective, and problematic in two main areas: (1) the criteria for determining which countries are “high risk” (and therefore requiring that seafood operations undergo labor audits to obtain MSC certification), and (2) the labor audit programs proposed for operations in countries considered to be high risk.

Assessing risk: An operator is exempt from a labor audit if the country they are operating in is deemed “low risk” by meeting at least two of the four indicators set out by MSC.[2] It is extremely problematic that operators seeking MSC certification are not all assessed on an individual basis but at a country level: if a country is considered low risk, then all seafood operators in that country are exempt from the labor requirements for the certification.

Risk assessment and classification on the basis of these criteria will effectively allow seafood operations that may have serious labor abuses to be certified without any labor due diligence. For example, if the current risk assessment criteria were applied to Taiwan, Hong Kong, South Korea, and the US, all would meet at least two of the criteria for low risk. This is alarming, given the well-documented occurrence of forced labor in the seafood and/or fishing sectors in these countries. Other countries where there is some known indication of risk would also be excluded from the labor audits, such as Japan and Singapore. The indicator based on UN convention ratifications is also inadequate for assessing risk in certain countries. For example, Thailand has ratified seven of the conventions (well above the required five), despite known forced labor in its seafood industry.

Labor audit programs: If countries are classified as high risk by the above risk assessment criteria, operators must undergo one of three third-party labor audit programs. The Thai Seafood Working Group has expressed our concern about the use of SA8000, Amfori, and SEDEX as the accepted programs. They have proven to be ineffective in other supply chain industries, such as apparel, palm oil, and cocoa, in some cases with tragic results. One such example is the case of Ali Enterprises, an SA8000-certified Pakistani apparel factory, in which over 250 workers died in a fire on September 11, 2012. The SA8000 auditors had failed to detect and raise the issue of fire hazards and lack of emergency exits. Across multiple industries, there are many examples of the failure of auditors to detect or report violations and of buyers to remediate those violations that are reported.  We hope that the seafood industry can learn from these experiences to prevent harm to workers and improve working conditions.

As has been shared with MSC, the principal reasons that third-party labor audit systems are ineffective include: (1) lack of transparency (findings are not disclosed to workers or the public); (2) lack of auditor expertise in labor and industrial relations; (3) weak audit methodologies; (4) lack of worker engagement/inclusion; (4) absence of credible complaints mechanisms; (5) conflict of interest (auditors paid for by the company acquiring certification); and importantly, (6) absence of enforceable mechanisms for remediation.

Forced labor cannot be easily identified or seen as an isolated problem; it is an accumulation of labor rights abuse. Rights violations – such as lack of a contract, withholding of wages, or retention of identity documents – can all be elements of forced labor. Most workers do not end up in situations of forced labor overnight, but are gradually coerced through a range of widespread practices. For these reasons, the snapshot of working conditions provided by an audit is insufficient. All of these issues jeopardize the credibility of the MSC certification and raise serious cause for concern.

Given the weakness of the labor requirements for MSC’s CoC Certification, and given that the requirements will not be revisited for another three years, members of the Thai Seafood Working Group would like to reiterate three recommendations for MSC and other environmental certification bodies considering incorporating labor provisions:

  1. All seafood operations, regardless of which country they are in, should be required to conduct proper due diligence on labor requirements for MSC certification. This change would reflect the challenges in ranking countries into low and high risk; the invisibility and vulnerability of seafood and fishing workers in many countries; and the importance of ensuring that child and forced labor are detected, prevented, and addressed throughout the supply chain.
  2. Develop a complaints mechanism that provides workers and labor rights stakeholders with a channel to raise complaints or grievances about labor issues. Such a mechanism would serve to surface labor violations that are not captured in the MSC program – either because they may be occurring in operations in countries that are exempt from the labor requirements, or they are not being detected through the labor audits. This would serve as an important step in addressing the enormous challenges for monitoring labor rights in the seafood supply chain.
  3. Undertake genuine consultation with labor organizations to develop future iterations of this and other certifications. Engaging and consulting trade unions in the process of developing labor rights criteria should be a key part of stakeholder engagement from the beginning. As MSC considers changes and improvements to the social criteria in their certifications in the future, there should be thorough and genuine consultation with – and a greater incorporation of recommendations from –labor experts, unions, and worker-led organizations.

While we understand that this was MSC’s first step in an effort to address labor abuse in seafood supply chains, we are deeply disappointed by the standard put forward. Indeed, a low-bar approach such as this one is a missed opportunity and a poor model for other sustainability programs seeking to address the rights and well-being of seafood workers.



Conservation International


Freedom Fund

Freedom United


Green America

International Labor Rights Forum

Humanity United

Human Rights Now

Human Rights Watch

Stop the Traffik Australian Coalition

Sustainability Incubator



[1] The new labor requirements were released as part of MSC’s revised Chain of Custody Standard, which is a traceability certification for on-shore seafood operators. It guarantees that products sold with the MSC label can be traced throughout the supply chain to a certified source and, as of March 2019, that the products were not made using forced or child labor. Once certified, the entity is able to sell seafood with the MSC-certified label; see

[2] These are the Country Risk Assessment Process for SA8000; the ITUC Global Rights Index; ratification of at least five of select UN conventions; and the U.S. Department of Labor List of Goods Produced by Child Labor or Forced Labor.

Posted: January 1, 1970, 12:00 am

Amos Toh is senior researcher, artificial intelligence and human rights, at Human Rights Watch. Before joining HRW, he served as legal advisor to the United Nations Special Rapporteur on the right to freedom of opinion and expression and was a clinical teaching fellow at the University of California (Irvine) School of Law. As legal advisor, he coordinated the Special Rapporteur’s research and advocacy on digital rights and internet freedom. Amos was formerly Counsel and Katz Fellow at the Brennan Center of Justice, where he worked on government surveillance reform and religious profiling issues. Amos is a graduate of New York University School of Law and the National University of Singapore School of Law.

Posted: January 1, 1970, 12:00 am

Outside the office of New York State Governor Andrew Cuomo, demonstrators protest against a law that bars the state from investing in companies that support boycotts of Israel, New York City, June 9, 2016.

© 2016 Mark Apollo/Pacific Press/LightRocket via Getty Images

Anti-Semitic acts are on the rise again in Europe. Amid an alarming increase in violent attacks against Jews in France, including several high profile murders, scores of Jews have decided to leave France for Israel. Some Jewish teenagers in Germany, tired of attacks and abuse, are following suit. There are similar concerns about antisemitism in the UK, with around 1 in 3 British Jews saying they are thinking about emigrating as a result.

For many Europeans, the hatred of Jews of the 1930s and 1940s may seem like ancient history. For Jews, especially for those in Germany, the trauma and collective memory hasn’t faded.

Governments like Germany’s are right to be concerned about the virulent cancer of anti-Semitism. But the joint CDU/CSU, SPD, FDP and Bündnis90/Die Grünen motion that passed the Bundestag recently that present boycotts of Israel as anti-Semitic is misplaced and the wrong way to combat anti-Semitism. The German government should reject it.

There’s no doubt that some anti-Semites have learned to use the term “Israel” or “Zionist” as a substitute for “Jews.” That should be called out. But it’s also true that legitimate criticism of Israeli state actions is sometimes wrongly tarred as anti-Semitic.

Anti-boycott measures, from the US to Israel, have often targeted people concerned about human rights abuses in illegal settlements in the Israeli-occupied West Bank. Twenty-seven US states have passed laws or policies that penalize businesses, organizations or individuals that engage in or call for boycotts of Israel. Human Rights Watch research has shown that many US states are using these laws to punish companies that refuse to do business with illegal West Bank settlements.

New York, for example, has published a list of 11 companies it is prohibited from investing in under a 2016 executive order issued by Governor Andrew Cuomo. The list includes companies that merely cut settlement ties. One, the Co-operative Group, a UK grocery, stopped doing business with suppliers known to buy produce from settlements, while saying it “remained committed to sourcing products from and trading with Israeli suppliers that do not source from settlements.” Another, Cactus Supermarkets of Luxembourg, suspended offering Israeli produce until suppliers demonstrated that the goods do not come from settlements, while continuing to offer other Israeli imports, said a statement by an activist group following negotiations with the company.

Human Rights Watch rejects all forms of anti-Semitism, is not part of the BDS movement and takes no position on boycotts of Israel. But years of our research has shown that it is not possible to do business in the settlements without contributing to or benefitting from human rights abuse and violations of international humanitarian law. The only way companies can meet their obligations under the UN Guiding Principles on Business and Human Rights is to stop operating in settlements. Anti-boycott laws punish companies that take such action in line with the international legal responsibilities and the German and European Union position on settlements.

The Israeli government has its own anti-boycott legislation. One year ago, authorities used a 2017 amendment to its Law of Entry that bans entry to those who call for boycotts of Israel to revoke the work visa of my Human Rights Watch colleague Omar Shakir. When we challenged the deportation order in court, the government highlighted his promotion of our research activities of businesses such as Airbnb in settlements and our recommendation that they halt their activities in settlements because of their harm to Palestinians’ rights. Last month, an Israeli court upheld the deportation order, claiming that our work on businesses in settlement constitutes a call to boycott Israel. We have appealed to Israel’s Supreme Court.

Everyone has the right under international human rights law to express their views through non-violent means, however abhorrent one might find them, including participating in boycotts. Authorities may restrict speech, but only under narrow and stringent conditions.

David Kaye, the UN special rapporteur on the promotion and protection of the right to freedom of opinion and expression, has stated that “Boycott…has long been understood as a legitimate form of expression, protected under Article 19(2)” of the International Covenant on Civil and Political Rights (ICCPR). In a review of anti-boycott laws in the US, Kaye concluded that the legislation “appears clearly aimed at combatting political expression” and that “economic penalties designed to suppress a particular political viewpoint” would not meet the conditions under the ICCPR for permissible restraints on speech.

In Germany, the term boycott evokes memories of the boycott of Jewish-owned shops in the 1930s. To equate that dark chapter with boycotts of Israel over its rights abuses is to trivialize our history. Activists worldwide use boycotts to challenge rights abuses and seek political change. Boycotts played key roles in the US struggle for African-American rights and in international campaigns against apartheid in South Africa and atrocities in Darfur.

Instead of pressing ahead with anti-boycott measures likely to restrict free expression and target those campaigning for human rights, German authorities should address resurgent anti-Semitism by investigating and punishing threats and violence against Jews and other minorities and by condemning intolerant speech by far-right politicians, and by educating people on the dangers of unchecked hatred. Officials could start, for example, by addressing head-on the anti-Semitic bullying in our schools. These measures are far more likely to be effective than restrictions on speech, which in the end will do nothing to combat hatred. 

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am



This submission examines the human rights implications of Artificial Intelligence (AI) [1] and other data-driven technologies in welfare benefits programs, such as cash and food assistance programs. Through a series of case studies, this submission explains how States delegate key welfare functions, such as determinations of eligibility and benefits levels, to automated decision-making models, some of which rely on data mining, machine learning and other processes or technologies typically associated with the field of AI. It also assesses how automated decision-making interferes with the rights to privacy and social security, and the obligations of States to guarantee the exercise of these rights without discrimination and undue private interference.



Overview of Applicable International Human Rights Law

The right to privacy

Article 17 of the International Covenant on Civil and Political Rights (ICCPR), which derives from Article 12 of the Universal Declaration of Human Rights (UDHR), establishes the right to “the protection of the law” against “arbitrary or unlawful interference” with one’s “privacy, family, home or correspondence.”

The Human Rights Committee has concluded that the prohibition against “arbitrary or unlawful interference” establishes a two-part test. First, interferences with privacy can take place only “in cases envisaged by the law.”[2] Under this requirement, States must “specify in detail” in relevant legislation “the precise circumstances in which such interferences may be permitted,” and ensure that decisions “to make use of such authorized interference must be made only by the authority designated under the law, and on a case-by-case basis.”[3] Second, for interferences to be non-arbitrary, the Committee has concluded that they must be “proportionate to the end sought, and ... necessary in the circumstances of any given case.”[4]


The right to social security

Article 9 of the International Covenant on Economic, Social and Cultural Rights (ICESCR) and Article 22 of the UDHR recognize the right of everyone to “social security, including social insurance.” The Committee on Economic, Social and Cultural Rights (CESCR) has concluded that this right establishes the obligation of States to ensure that eligibility criteria for social security benefits are “reasonable, proportionate and transparent.”[5] Furthermore, the “withdrawal, reduction or suspension of benefits should be “based on grounds that are reasonable, subject to due process, and provided for in national law.”[6]

Access to information is also a precondition of the enjoyment of the right to social security. The CESCR has found that beneficiaries of social security schemes “must be able to participate in the administration of the social security system.”[7] Accordingly, the system should “ensure the right of individuals and organizations to seek, receive and impart information on all social security entitlements in a clear and transparent manner.”[8] 


Non-Discrimination Obligations

Article 2(1) of the ICCPR and Article 2(2) the ICESCR require States to guarantee Covenant rights without discrimination of any kind based on “race, color, sex, language, religion, political or other opinion, national or social origin, property, birth or other status.” Article 26 of the ICCPR additionally guarantees the right “to all persons equal and effective protection against discrimination on any ground.” The Human Rights Committee has found that Article 26 establishes an “autonomous right” that “prohibits discrimination in law or in fact in any field regulated and protected by public authorities.”[9]

In the context of social security, the CESCR has found that States should ensure that social security schemes “do not discriminate in law or in fact.”[10] States should also “pay special attention” to those that disproportionately experience difficulties in accessing social security, such as women, people with disabilities, minorities and “casual” or “seasonal” workers.[11]        


Obligation to Protect Against Third-Party Interference

States have a duty to protect individuals from undue third-party interferences with their rights to privacy and social security. The Human Rights Committee has concluded that Article 2(1) of the ICCPR, which establishes the State’s duty to “ensure” the right to privacy and other Covenant rights, imposes a positive obligation to protect individuals against “acts committed by private persons or entities that would impair the enjoyment of [these] rights.”[12] This duty requires the adoption of “appropriate measures” or “due diligence to prevent, punish, investigate or redress the harm caused by ... private persons or entities.”[13]

The CESCR has categorized State obligations under the ICESCR as obligations to respect, protect and fulfill Covenant rights.[14] The obligation to protect the right to social security requires States to prevent corporations and “agents acting under their authority” from interfering with that right.[15] In the context of social security schemes that are “operated or controlled by third parties,” States “retain the responsibility of administering the national social security system and ensuring that private actors do not compromise equal, adequate, affordable, and accessible social security.”[16] To prevent abuses, “an effective regulatory system must be established which includes framework legislation, independent monitoring, genuine public participation and imposition of penalties for non-compliance.”[17]


Human Rights Implications of Using Technology in Cash and Food Assistance Programs

States rely on AI and related technologies to automate two critical stages of the welfare distribution process: the verification of claimants’ identity, and the assessment of eligibility and benefits levels. Throughout the entire welfare delivery cycle, States also employ these technologies to investigate, adjudicate and impose penalties for fraud.


Identity Verification

Aadhaar (India)

The use of AI to verify the identity of welfare recipients may be part of a broader push to establish national digital identity frameworks that manage individuals’ access to government entitlements through a single, government issued identity. In 2009, India launched Aadhaar, a digital identity framework that assigns every citizen a unique twelve-digit identification number linked to the individual’s biometric and demographic data. Under the Aadhaar Act of 2016, beneficiaries of various government welfare programs such as the Public Distribution System (PDS), which provides subsidized food grains to millions of households, are required to register and use Aadhaar to access their entitlements.[18]

In 2018, the Indian Supreme Court upheld the government’s authority to mandate Aadhaar as a precondition for accessing food rations and other welfare benefits.[19] However, the Court ruled that certain provisions of the Aadhaar Act were unconstitutional, and also barred the private sector from seeking access to Aadhaar data.[20] In February 2019, the government passed amendments to the Aadhaar Act that restored such access and bypassed the Court’s ruling.[21]


Human Rights Watch has found that eligible families have been denied access to subsidized food grains and other benefits because they did not have an Aadhaar number, had not linked it to their ration cards or experienced failures in authenticating their fingerprints.[22] Authentication failures disproportionately affect manual laborers, older persons and other individuals with worn fingerprints.[23] Since the Aadhaar machines installed in food distribution outlets require an internet connection, poor connectivity in rural areas has also led to disruptions in food distribution schedules.[24] Local activists have found that Aadhaar-related denials of food rations have led some to starve to death.[25]

Aadhaar also imposes invasive biometric identification and data collection requirements as conditions for accessing subsidized food grains and other essential public services. These requirements have created the world’s largest database of biometric identity information, escalating the risk of unnecessary and disproportionate surveillance.[26] To mitigate this risk, the Supreme Court imposed several restrictions, including the requirement of judicial approval for law enforcement access to Aadhaar data, and a six-month limit on the retention of authentication records and transaction logs.[27] However, these changes do not address the scope of biometric data and personal information collected under the program. Human Rights Watch has also raised concern about the multiple data breaches associated with Aadhaar since its implementation.[28]

These interferences with privacy disproportionately affect minorities: for example, local activists fear that transgender individuals are at greater risk of discrimination and persecution when they are forced to disclose their gender identity to the government, or if such information is leaked to the public.[29] These risks also raise the possibility that transgender individuals will be deterred from seeking access to essential public services linked to Aadhaar.  


Knowledge Based Authentication System (California, United States)

Countries without national ID schemes also rely on automated decision-making to verify the identity of welfare claimants by comparing multiple sources of identity-related information drawn from a wide range of government and private databases. In the United States, the California state legislature in 2017 amended the Welfare and Institutions Code to replace fingerprint imaging with an “automated, nonbiometric” method for verifying the identity of applicants to the CalWORKs program, which provides cash assistance to needy families.[30] The Department of Social Services (DSS) selected US-based private company Pondera Solutions[31] to conduct a pilot of a cloud-based identity verification system known as the Knowledge Based Authentication system (KBA).[32] The pilot was conducted in six counties.[33]

KBA checks a CalWORKs application against “over 10,000 public sources” of data from “dozens of categories and hundreds of jurisdictions,” including data from credit bureaus, government agencies and “utility and telephone companies.”[34] This initial assessment is “designed to verify that the identity provided to the program is legitimate.”[35] It also generates a multiple-choice quiz for applicants that “seeks to ensure that the applicant is in fact the individual that they are representing themselves to be.”[36] Applicants are assigned a fraud risk code based on the system’s initial assessment and their answers to the quiz.[37]

At the conclusion of the pilot in October 2017, DSS announced that it was intending to roll out KBA for phone or online benefits applications by the summer of 2018.[38] It is unclear, however, whether it has adhered to this timeline.


In its report on the results of the pilot, DSS did not explain how the KBA analyzes the wide variety of data sources at its disposal to verify an applicant’s identity and generate quiz questions. Although KBA standardizes the spelling of addresses “to avoid misspellings and other common mistakes,” it is unclear how the system responds to other errors in the data it is provided, such as discrepancies in dates, phone numbers and demographic details.[39] The report also acknowledges KBA’s potential to generate “false positives,” but does not provide information about DSS’s plans to prevent or mitigate such errors.[40] 

This lack of transparency makes it difficult for welfare claimants and the broader public to assess the reliability, accuracy or fairness of KBA’s risk assessment calculus. If incorrect information is associated with claimants and answers to the quiz questions are wrongly marked as errors, it will be difficult for them to identify the source of the error and hold the relevant authorities accountable. In addition, KBA’s analysis of large datasets containing a wide range of sensitive and personal information raises questions about how the system safeguards applicants’ privacy.

The Coalition of California Welfare Rights Organizations has also raised concern that KBA’s multiple-choice quiz creates additional obstacles for marginalized populations. Families that have been homeless for a long time may be unable to answer questions such as “How long have you lived in your current residence?” or “Which of the following streets have you ever lived or used as your address?”[41] Furthermore, questions regarding residential and relationship history in the United States assume that respondents have longstanding community ties, and are ill-suited to the needs and concerns of newly arrived immigrant families.


Assessment of Eligibility and Benefits Levels 

Ontario Works (Ontario, Canada)

Governments are also replacing or supplementing case workers’ assessments of eligibility and benefits levels with predictive analytics and other AI-based assessment tools. Since November 2014, Ontario Works, the financial assistance program of the Canadian province of Ontario, has been relying on the Social Assistance Management System (SAMS) to automatically generate decisions on eligibility for cash transfers and other benefits. Decisions are generated based on data that frontline workers collect from applicants and recipients and subsequently “fit into narrow-drop down menu categories.”[42]

SAMS is based on Cúram, a customizable off-the-shelf software sold by IBM as a platform for “complete intake, eligibility determination and benefit calculation for social programs.”[43] Latest versions of the software are also equipped with functions to monitor impermissible instances of “concurrent eligibility” in food and cash assistance programs, potentially indicating the system’s ability to perform both benefits assessments and fraud detection.[44] Cúram is also used to administer welfare programs in Alberta, North Carolina, Hamburg, Queensland and New Zealand.[45]

A 2015 audit of SAMS conducted by the state’s Auditor General found that SAMS suffered from “serious defects and was not fully functional,”[46] leading to potential underpayments of benefits totaling $51 million CAD.[47] In one case, SAMS erroneously deducted $32 from a client’s total benefit payments each month after it incorrectly determined that the client had been previously overpaid.[48]

The Auditor General also found that SAMS “automatically generated” letters to beneficiaries with “incorrect information” that caused “stress and confusion.”[49] For example, a letter sent to two beneficiaries, whom owed $1,328 in overpayments, accused them of owing $8,736.[50] Another letter notified a beneficiary that their benefits would be withdrawn because they no longer lived in Ontario, but caseworkers found that the beneficiary “had never left Ontario.”[51] 


Universal Credit (United Kingdom)

The Special Rapporteur on extreme poverty and human rights has observed that the Universal Credit (UC), the UK’s welfare benefits program, “is only possible because of the automated calculation of benefits.”[52] The Real Time Information (RTI) system calculates UC payments based on earnings information reported by employers to Her Majesty's Revenue and Customs (HMRC), the country’s tax authority.[53]

It appears that RTI’s calculations are unable to correct for late, inaccurate or missing reports, and this can lead to delays and errors in UC payments.[54] In the 2016/2017 fiscal year, 5.7% of 590m UC payments were marred by late reporting.[55] To address RTI’s inability to take into account reporting errors, HMRC and the Department of Work and Pensions, which oversees UC, have established a special joint initiative known as the Late, Missing and Incorrect RTI Project.[56]



The failures of automated decision-making in Ontario Works and UC indicate that governments have overestimated the technology’s capacity to conduct complex and context-sensitive assessments of eligibility and benefits levels. Benefits calculators are only as accurate as the data provided to them: unlike case workers, these systems are unable to investigate the reasons for data inaccuracies and other discrepancies and make necessary adjustments on a case-by-case basis.

Despite these limitations, both programs fail to supplement automated decision-making with human review that ensures benefits changes are reasonable and in accordance with due process and domestic law requirements. In the Ontario Works program, Professor Jennifer Raso of the University of Toronto Law School has found that the design of SAMS “obstructs frontline workers from challenging the substance of its decisions.”[57] For example, SAMS does not permit frontline workers to challenge its assessments of whether Ontario Works recipients with a history of living in the same household are dependent on each other – a common ground for reducing the value of benefits.[58] 

In the UK, service workers have not been provided with training that enables them to effectively troubleshoot RTI and other IT errors that may lead to the withdrawal, reduction or suspension of benefits. According to welfare researchers from the University of Birmingham and the University of Leeds, former staff of Jobcentre Plus, an agency which administers UC’s benefits for jobseekers, “described being permanently on the ‘back foot’, in that digital services were rolled out without staff being given the relevant training.”[59] A former UC call center worker in Grimsby also told The Guardian that there was “massive variation” in staff’s understanding of the UC policies and systems, leading to contradictory responses to the same query.[60] 


Detection, Investigation and Punishment of Welfare Fraud

Systeem Risico Inventarisatie (The Netherlands)

Governments increasingly rely on automated fraud detection systems to detect and flag risks of welfare fraud. In the Netherlands, the Ministry of Social Affairs and Employment operates the Systeem Risico Inventarisatie or System Risk Indication (SyRI), which is used by several municipal governments to detect benefits fraud. SyRI flags individuals as potential fraud risks through an algorithmic risk assessment tool that draws on multiple sources of data, including tax returns.[61] However, the government has offered few details on the specific types of data used and the criteria for determining risk. It has rejected calls for transparency about how the algorithm works, claiming that disclosing such information would reduce its effectiveness in detecting fraud.[62] It has also not explained the circumstances under which case workers or fraud investigators may deviate from these risk assessments, if at all. 


Online Compliance Intervention (Australia)

Several governments do not only automate assessments of welfare fraud, but also the imposition of penalties such as fines or the reduction or withdrawal of benefits. In July 2016, Australia’s Department of Human Services (DHS) launched Online Compliance Intervention (OCI), a fully automated income data verification system that generates debt notices based on differences between fortnightly income figures reported by welfare beneficiaries and their employers.[63] A Parliamentary inquiry conducted by the Senate’s Community Affairs Reference Committee found that OCI’s income calculation formula is not programmed to take into account fluctuations in a beneficiary’s income, leading to “inaccurate calculations of debt,” particularly for casual or seasonal workers with irregular incomes.[64] OCI is also unable to make adjustments for employer error.[65] 

DHS does not require manual review of the OCI’s findings, instead placing the onus on beneficiaries to submit evidence rebutting debt notices.[66] However, affected beneficiaries and their representatives testified before the Senate that debt notices either provided inadequate information or were too complicated to understand, making them difficult to challenge.[67] Some complained that they had to submit Freedom of Information requests to compel DHS to release information about how their debts were calculated.[68] 

The Senate Committee has urged DHS to put OCI “on hold” until these issues are resolved,[69] but the Department has rejected this recommendation and claimed that its implementation has gone “quite well.”[70]


Michigan Integrated Data Automated System (Michigan, United States)

In October 2013, Michigan’s Unemployment Insurance Agency (UIA) launched the Michigan Integrated Data Automated System (MiDAS) to adjudicate and impose penalties for unemployment benefits fraud. Between October 2013 and August 2015, MiDAS was programmed to automatically treat differences between income figures reported by beneficiaries and their employers as evidence of fraud.[71] The system was not capable of investigating whether there are legitimate reasons for these discrepancies, such as employer error or pay disputes.[72] Like OCI, MiDAS was also unable to determine whether these discrepancies are attributable to fluctuations in a beneficiary’s income.[73]  

Based on its initial assessments, MiDAS sent beneficiaries suspected of fraud online multiple-choice questionnaires asking whether they are “intentionally provid[ing] false information to obtain benefits you were not entitle[d] to receive,” and “[w]hy ... you believe you were entitled to benefits.”[74] Failure to respond in ten days, or a response that MiDAS deemed unsatisfactory, would automatically trigger conclusive determinations of fraud.[75] Based on these determinations, MiDAS would terminate the benefits of affected beneficiaries and initiate proceedings to seize their tax refunds or garnish their wages.[76]

UIA subsequently found that, between October 2013 and August 2015, about 44,000 of the 62,784 determinations of fraud that MiDAS generated were in error.[77] In a lawsuit that a group of beneficiaries filed against UIA, the U.S. Federal Court of Appeals for the region concluded that MiDAS “did not allow for a fact-based adjudication or give the claimant the opportunity to present evidence to prove that he or she did not engage in disqualifying conduct.”[78]

Despite these failures, UIA continues to operate MiDAS.[79] It has committed to additional data analysis to detect benefits payments needing “further review” and to enhance the appeals process.[80] However, it is unclear whether UIA has made any changes to the underlying data matching algorithm or incorporated meaningful human review into the system’s fraud detection functions.


Automated Verification of Job Activity Reports (Sweden)

The Swedish Employment Service (Arbetsförmedlingen) relies on automated decision-making to verify whether recipients of unemployment benefits have complied with job-seeking and other workfare obligations, and issue warnings, withhold payments and enforce other sanctions based on these assessments.[81] At the end of 2018, Arbetsförmedlingen discovered a 10 – 15% error rate in the automated verification of beneficiaries’ job activity reports, potentially leading to 70,000 erroneous decisions to withhold benefit payments.[82] These errors have forced Arbetsförmedlingen to manually screen all job activity reports until the system can be repaired.[83] 



These cases reinforce the need for appropriate human review that corroborates fraud findings generated by automated systems, as well as clear, transparent and accessible appeals mechanisms that enable beneficiaries to meaningfully challenge these findings. Without these safeguards, limitations or errors in automated decision-making potentially lead to mass violations of the right to social security. The Swedish and Michigan examples illustrate that the automation of fraud determinations at scale replicates errors in data processing and analysis across the entire system, leading to incorrect benefits changes and penalties that affect thousands of beneficiaries. The lack of transparency compounds these failures, preventing beneficiaries from accessing information about their case or participating in its adjudication. 

These failures also illustrate the potential for welfare discrimination based on beneficiaries’ socio-economic backgrounds. Flaws in OCI and MiDAS’ income calculation formulae, for example, disproportionately affect workers with irregular incomes, whom the CESCR has designated as a protected category in the social security context.[84] The Michigan Law School Unemployment Insurance Clinic has also raised concern that beneficiaries experiencing financial hardship have extremely limited options to challenge MiDAS’s determinations: Charges of fraud disqualify them from free representation under the state’s pro bono program, and it is unlikely that they are able to afford private representation.[85] Under OCI, the Senate Committee heard evidence that beneficiaries with poor literacy or English language skills found it particularly difficult to understand the highly technical language used in debt notices.[86]


The Role of the Private Sector

The case studies outlined in this submission show that the private sector is key in developing and operating automated systems of welfare governance. Companies involved range from those providing specialized fraud detection services to large enterprise software manufacturers. However, it is unclear whether these companies have established policies or processes that meaningfully address their human rights impacts.

These public-private partnerships make it difficult to hold both State and non-State actors accountable for failures in the welfare delivery services that are outsourced. AI Now, an organization dedicated to examining AI’s public and social impacts, has found that risk assessment models and other automated decision-making tools are typically hidden behind broad assertions of intellectual property and trade secrets, making it difficult for affected rights holders and the broader public to scrutinize their potential for discrimination and other human rights impacts.[87] There also does not appear to be much pressure on companies to conduct human rights impact assessments or consultations with welfare recipients during the design, customization and implementation of welfare delivery software, particularly since governments are not insisting on adherence to the UN Guiding Principles.

Implementation of the UN Guiding Principles in the Information and communications technology sector, which has hitherto focused on the responsibilities of internet and telecommunications companies to respect freedom of expression and privacy, offers general guidance and best practices that are adaptable to the commercial delivery of welfare-related services.

Human rights due diligence is a central component of these responsibilities, and requires impact assessments that address issues of privacy, discrimination and exclusion early on in the design and engineering phase, internal training, dialogue and collaboration on these issues, and regular consultations with civil society and affected rights holders.[88] Companies should also establish meaningful transparency measures, such as policies to disclose the outcomes of impact assessments and the concrete steps they have taken to prevent or mitigate human rights risks.[89] Furthermore, companies have a responsibility to provide access to effective remedies (such as financial restitution) when they have “caused or contributed to adverse [human rights] impacts.”[90]

In the context of digital welfare, companies should, at a minimum, provide accessible explanations of how AI and other data-driven technologies are integrated into welfare decision-making, disclose and address automation errors in a timely fashion, submit to audits of algorithms and training data by external assessors, and develop processes for identifying, correcting and mitigating discrimination and bias in system inputs and outcomes.[91] 

In accordance with their obligations to protect against private interference with Covenant rights, States should establish implementation of the UN Guiding Principles as a mandatory condition for the sale of identity verification, benefits assessment and fraud detection products and services to welfare agencies and other relevant authorities.




[1] The UN Special Rapporteur on the right to freedom of opinion and expression has defined Artificial Intelligence (AI) as the “constellation of processes and technologies enabling computers to complement or replace specific tasks otherwise performed by humans, such as making decisions and solving problems.” Report of the Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression, David Kaye, Human Rights Council, A/73/348, Aug 29, 2018, ¶ 3 (internal quotation marks omitted),

This is also the definition that Human Rights Watch uses in this submission.

[2] Human Rights Committee, General Comment 16: Article 17 (Right to Privacy), ¶ 3,

[3] Id., ¶ 8.

[4] See e.g. Toonen v. Australia, Communication No. 488/1992, U.N. Doc CCPR/C/50/D/488/1992 (1994), ¶ 8.4,; Antonius Cornelis Van Hulst v. Netherlands, Communication No. 903/1999, U.N. Doc. CCPR/C/82/D/903/1999 (2004), ¶ 7.6,

[5] CESCR, General Comment 19: The right to social security (art. 9), Nov 23, 2007, ¶ 24,

[6] Id., ¶ 24.

[7] Id., ¶ 26.

[8] Id.

[9] Human Rights Committee, General Comment No. 18: Non-discrimination, Nov 10, 1989, ¶ 12,

[10] CESCR, General Comment 19, supra n. 5, ¶ 30.

[11] Id., ¶ 31.

[12] Human Rights Committee, General Comment No. 31 [80]: The Nature of the General Legal Obligation Imposed on States Parties to the Covenant, Mar 29, 2004, ¶ 8,

[13] Id.

[14] CESCR, General Comment 19, supra n. 5, ¶ 43.

[15] Id., ¶ 45.

[16] Id., ¶ 46.

[17] Id.

[18] “India: Identification Project Threatens Rights”, Human Rights Watch news release, Jan 13, 2018, (“HRW Jan 13 release”).

[19] “India: Top Court OK’s Biometric ID Program”, Human Rights Watch news release, Sep 27, 2018,

[20] Id.

[21] Gautam Batia, The Aadhaar ordinance raises serious constitutional concerns, Hindustan Times, Mar 1, 2019,

[22] HRW Jan 13 release, supra n. 18.

[23] Id.

[24] Jean Drèze, Nazar Khalid, Reetika Khera, Anmol Somanchi, Aadhaar and Food Security in Jharkhand

Pain without Gain?, Economic & Political Weekly (Dec 16, 2017 Vol. LII No. 50), 55,

[25] “Of 42 'Hunger-Related' Deaths Since 2017, 25 'Linked to Aadhaar Issues'”, The Wire, Sep 21, 2018,

[26] HRW Jan 13 release, supra n. 18.

[27] Justice K.S. Puttaswamy (Retd.) et al. v. India et al., Writ Petition (Civil) No. 494 of 2012 & connected matters, majority opinion by A.K. Sikri, J, ¶¶ 205, 344, 345,

[28]  HRW Jan 13 release, supra n. 18.

[29] “Aadhaar exposes transgenders to violence, discrimination and surveillance: SC told”, India Today, May 20, 2018,

[31] “Analytics With Context”, Pondera Solutions, Human Rights Watch contacted Pondera to obtain more information about the KBA but it did not respond at the time of the submission.

[32] California Department of Social Services, Summary of Options for

Replacing the Statewide Fingerprint Imaging System (October 2017),, 5 (“CDSS 2017 report”).

[33] Id., 5.

[34] CDSS 2017 report, supra n. 32, 18.

[35] Id.,

[36] Id., 12 – 13. 

[37] Id., at 15; Kathleen Wilson, Ventura County joins ID program for welfare benefits (Mar 26, 2017),

[38] CDSS 2017 report, supra n. 32, 9.

[39] Id., 18.

[40] Id., 17.

[41] “Advocate Response to DSS Options for Replacing the Statewide Fingerprint Imaging System (SFIS)”, Coalition of California Welfare Rights Organizations, Inc., Dec 4, 2017, 12,  

[42] Jennifer Marie Raso, Administrative Justice: Guiding Caseworker Discretion, 2018, 245 – 246, (“Raso thesis”).

[45] “Ministry of Community and Social Services SAMS Transition Review”, Apr 30, 2015, 60 – 62,

[46] “SAMS—Social Assistance Management System,” 2015 Annual Report of the Office of the Auditor General of Ontario, Fall 2015, 479,

[47] Id., 485.

[48] Id., 480.

[49] Id., 481.

[50] Id. 

[51] Id.

[52] “Statement on Visit to the United Kingdom, by Professor Philip Alston, United Nations Special Rapporteur on extreme poverty and human rights”, OHCHR News and Events, Nov 16, 2018,

[54] “We know that Universal credit is a mess, but what about HMRC Real Time Information system?”, Disabled People Against Cuts, Oct 25, 2017,; Patrick Butler,  Universal credit IT system ‘broken’, whistleblowers say, The Guardian, Jul 22, 2018,

[56] Response to Sept 4, 2017 FOI request, DWP Central Freedom of Information Team, FOI Ref No. 3665, Sept 26, 2017,

[57] Raso thesis, supra n. 42, 254.

[58] Id., 254 – 255.

[59] Kayleigh Garthwaite, Jo Ingold, and Mark Monaghan, Universal Credit and the perspectives of ex-Jobcentre Plus staff, LSE British Politics and Policy blog, Jan 15, 2019,

[60] Patrick Butler, Universal credit staff: 'It was more about getting them off the phone', The Guardian, Jul 22, 2018,

[61] “Automating Society: Taking Stock of Automated Decision Making in the EU”, AlgorithmWatch and Bertelsmann Stiftung, January 2019 (1st ed.), 101,

[62] “Beantwoording Kamervragen over 'Algoritme voorspelt wie fraude pleegt bij bijstandsuitkering'”, Apr 12, 2018,

[63] The Senate, Community Affairs References Committee, “Design, scope, cost-benefit analysis, contracts awarded and implementation associated with the Better Management of the Social Welfare System initiative,” 21 June 2017, 13, (“Senate Committee OIC Report”).

[64] Id., 32, 34.

[65] Id., 32.

[66] Id., 15.

[67] Id., 57 – 58.

[68] Id.

[69] Id., ix.

[70] Rae Johnston, The Department Of Human Services Says RoboDebt 'Went Well' And 'Delivered Lots Of Savings', Gizmodo, Mar 27, 2018,

[71] Cahoo, et al. v. SAS Analytics Inc., et al., Nos. 18-1295/1296 (6th Cir. 2019), 3,

[72] Id.

[73] Id.

[74] Id., 3 – 4.

[75] Id., 4 – 5.

[76] Id.

[77] “Michigan's unemployment agency completes review of fraud determination cases; comprehensive changes underway to improve customer service and operations”, press release,,4669,7-192-47796-428651--,00.html; Paul Egan, After Falsely Accusing Thousands of Unemployment Fraud and Wrongly Taking Their Money, Michigan Makes Amends. But Is It Enough?, Governing, Aug 15, 2017,

[78] Cahoo v. SAS Analytics, supra n. 71, 3.  

[79] “Final Agency Response for OAG Performance Audit of MiDAS System,” State of Michigan Department of Talent and Economic Development, Jun 17, 2016,

[80] Id., 6.

[81] Tom Wills, Sweden: Erroneous algorithm stops payments for over 70,000 unemployed, AlgorithmWatch, Feb 28, 2019,

[82] “SVT avslöjar: Datafel kan ha skapat tiotusentals felaktiga beslut hos Arbetsförmedlingen”, SVT Nyheter, Feb, 13, 2019,

[83] Id.

[84] CESCR, General Comment 19, supra n. 5, ¶ 31.

[85] Steven Gray, The Future of UIA Claims After the MiDAS Fraud Scandal, University of Michigan Law School Unemployment Insurance Clinic, Jan 15, 2015, 5,

[86] Senate Committee OIC Report, supra n. 63, 51 – 54.

[87] Dillon Reisman, Jason Schultz, Kate Crawford, Meredith Whittaker, Algorithmic Impact Assessments: A Practical Framework for Public Agency Accountability, AI Now, April 2018, 13,

[88] Report of the Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression, David Kaye, Human Rights Council, A/HRC/35/22, Mar 30, 2017, ¶¶ 52 – 60,

[89] Id., ¶¶ 70 – 72.

[90] Id., ¶¶ 73 – 75.

[91] A/73/348, supra n. 1, ¶¶ 65 – 70.

Posted: January 1, 1970, 12:00 am

People wait at a a cable car station decorated with graffiti of a Georgian miner in the city of Chiatura, Georgia, August 31, 2018.

© 2019 AP Photo/Evgeniy Maloletka

Protesting poor working conditions and demanding higher wages, about 150 workers at a manganese mine in Chiatura, Georgia, went on strike May 15. On Friday, 15 of the workers reportedly began a hunger strike. Manganese is an element used in making stainless steel.

Human Rights Watch conducted research on the working conditions in extractive industries in Georgia, and heard how miners, like those on strike, work shifts of 15 consecutive days, working 12 hours a day underground, including at night. During this time, they are required to live in a company dormitory, where, some told us, food is insufficient for the effort required on the job, and restrictions on their movement keep them away from their homes and families nearby.

Negotiations between workers and the mine operators appear to be temporarily stalled.

In April, the United Nations Working Group on Business and Human Rights expressed concern that the organization of work and the dormitory system at the manganese mine were not consistent with international labor standards.

Poor working conditions are endemic in Georgia, where labor laws and enforcement remain substandard – even after recent government reforms. Georgian law does not provide for mandatory weekly rest or breaks during shifts, and the country’s labor inspectorate is not permitted to check all working conditions. In March, the European Committee of Social Rights noted that the lack of an appropriate authority to monitor daily and weekly working limits places Georgia in violation of its obligations under the European Social Charter.

The manganese company should listen carefully to workers’ complaints and take steps to ensure safe and fair working conditions. Georgia should bring its labor code in line with International Labour Organization standards and put in place a full labor inspectorate with powers and resources to monitor and enforce all labor rights.  

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am

Equatorial Guinea's President Teodoro Obiang Nguema Mbasogo addresses the 73rd session of the United Nations General Assembly, at U.N. headquarters, Thursday, Sept. 27, 2018.

© 2018 AP Photo/Richard Drew

The United Nations’ Universal Periodic Review (UPR) process is meant to give countries a chance to scrutinize the human rights records of fellow UN member states. But countries with awful human rights records often respond to the scrutiny by resorting to the same script: claim to be a paragon of virtue, while in reality being an egregious rights abuser.

Such hypocrisy was on full display this week, as the spotlight was on Equatorial Guinea, an oil-rich central African country led by the world’s longest-serving president. States raised a litany of government abuses – harassment of human rights defenders, lack of an independent judiciary, systemic corruption, and abject poverty despite having the highest per-capita income in Africa. And, as per the script, Equatorial Guinea’s government representative, Alfonso Nsue Mokuy, dismissed the criticisms, absurdly claiming, for example, that the country is an “open democracy” and “there is no violence against women.”

But that didn’t stop Mokuy from accepting nearly all recommendations aimed at addressing the rights abuses he previously denied, including eliminating poverty and combatting corruption. Acknowledging “challenges ahead,” he promised a “new paradigm of respect for human rights and fundamental freedoms.”

There is every reason to believe these are empty words, designed to mask rather than address decades of callous neglect for the basic rights of Equatorial Guineans.

But what would happen if Equatorial Guinea actually made good on these promises?

Imagine if the government respected the right of all Equatorial Guineans to express themselves freely, instead of harassing, beating, or arresting those who defend human rights and promote good governance. The press and civil society would be empowered to act as a check on arbitrary detention, torture, corruption, and other abuses.

Imagine if there were truly independent investigations of corruption at the highest levels, including the vice president, who is also the president’s son and was convicted in absentia by a French court for laundering over $120 million. It would ensure the nation’s oil wealth actually benefitted the public instead of a small political elite.

Imagine if that wealth was used to improve health and education, and access to water and food. Over half of Equatorial Guinean children under four do “not have access to adequate food,” and over half who are primary-school aged were not in school, according to UN agencies. Their lives would be transformed.

Governments should hold Equatorial Guinea accountable for implementing the recommendations made at the UPR – making this dream one step closer to reality.

Author: Human Rights Watch
Posted: January 1, 1970, 12:00 am