Dutch government continued SyRI welfare fraud profiling after ban
The Dutch government deployed the SyRI algorithm to profile welfare recipients in low-income neighborhoods, despite a 2020 court ruling that the system violated EU human rights. The algorithm flagged thousands of parents for fraud based on illegal risk factors such as dual nationality, leading to false accusations, bankruptcy, and trauma. The government has announced an external review of the practices.
- Company involved
- Ministry of Social Affairs of the Netherlands
- AI system involved
- SyRI
10 source articles · read the reporting →
Suzhou clarifies civility scoring system after public backlash
The Suzhou municipal government announced a civility code system that assigns scores to residents based on their behavior, including traffic violations and volunteer work. The system, currently in a trial phase, aims to generate a personal portrait for each resident. Following public backlash and concerns about privacy and potential abuse, an official stated that participation will not be mandatory and that a bad score will not affect citizens. The system is still under development.
- Company involved
- Suzhou municipal government
- AI system involved
- Civility code
10 source articles · read the reporting →
Momus Analytics' jury-selection algorithm criticised for racial bias
Momus Analytics, a legal technology company, uses machine learning to rank potential jurors for attorneys. Experts and researchers who reviewed the company's patent application say the algorithm relies on demographic characteristics such as race, which may violate constitutional prohibitions against discriminatory jury selection. The company claims its software has helped secure verdicts worth over $940 million, but did not respond to requests for comment.
- Company involved
- Momus Analytics
- AI system involved
- Momus
9 source articles · read the reporting →
Polish government suspends culture fund after algorithm grants millions to wealthy stars
The Polish government used an algorithm to allocate 400 million zloty in COVID-19 culture support funds. The algorithm awarded large sums to high-earning star performers, including 1.89 million zloty to a company managed by Paweł Golec. After public criticism, Culture Minister Piotr Gliński suspended the fund and announced a review. The government defended the algorithm, stating it was based on financial criteria and aimed to protect jobs.
- Company involved
- Polish Ministry of Culture and National Heritage
9 source articles · read the reporting →
Allstate's proposed retention model charged big spenders more in Maryland
Allstate proposed an auto insurance rate adjustment plan in Maryland using a retention model algorithm that varied premium changes based on existing premiums. The algorithm would have increased rates up to 20% for customers already paying the highest premiums while capping increases at 5% for others, and would have denied meaningful discounts to thousands owed reductions. The Maryland Insurance Administration rejected the plan as discriminatory, but Allstate claims it was withdrawn and has continued proposing similar models elsewhere. The proposal never took effect in Maryland.
- Company involved
- Allstate Corporation
- AI system involved
- retention model
9 source articles · read the reporting →
New York City's McKinsey-led jail violence program manipulated data, violence increased
New York City paid McKinsey & Company $27.5 million to reduce violence at Rikers Island jail complex. McKinsey designed a predictive algorithm called the Housing Unit Balancer and Restart housing units, but jail officials and McKinsey consultants stacked the units with compliant inmates to artificially lower violence numbers. Violence actually increased by nearly 50% during the project. The city eventually decided to close Rikers.
- Company involved
- New York City Department of Correction
- AI system involved
- Housing Unit Balancer (HUB)
10 source articles · read the reporting →
Portland Metro ends Replica partnership over data privacy concerns
Portland Metro, an elected regional government in Oregon, ended its pilot project with movement data company Replica after a disagreement about data sharing. Portland Metro requested raw, disaggregated data, which Replica refused to provide, citing user privacy concerns. The partnership was terminated without payment.
- Company involved
- Portland Metro
- AI system involved
- Replica
10 source articles · read the reporting →
LA's VI-SPDAT housing scoring system gives lower priority to Black and Latino people
The Los Angeles Homeless Services Authority uses the VI-SPDAT scoring system to prioritise unhoused people for subsidised permanent housing. An investigation by The Markup found that Black and Latino people consistently receive lower vulnerability scores than White people, leading to lower priority for housing. The agency has acknowledged the racial disparities and is working on a new tool, but continues to use the current system.
- Company involved
- Los Angeles Homeless Services Authority
- AI system involved
- VI-SPDAT
10 source articles · read the reporting →
Serbia's Social Card law could harm marginalized groups
Serbia's Social Card law, which entered into force on 1 March 2022, establishes a centralized government database to assess eligibility for social security support. Amnesty International and seven other rights organizations submitted a legal opinion to the Constitutional Court, alleging that the automated system is an intrusive surveillance system that could discriminate against Roma communities and people with disabilities. The system processes 130 categories of personal data and lacks transparency, potentially leading to errors and denial of benefits.
- Company involved
- Serbian government
- AI system involved
- Social Card system
10 source articles · read the reporting →
Nevada's AI grad score model cuts funding for low-income students
Nevada replaced its income-based school funding formula with a machine learning model developed by Infinite Campus that generates a 'grad score' for each student. The model reduced the number of students eligible for supplemental funding from 288,000 to 63,000, disproportionately affecting low-income students. Critics argue the model lacks transparency and de-emphasizes economic status, potentially leaving many high-poverty schools underfunded.
- Company involved
- Nevada Department of Education
- AI system involved
- Grad score model
7 source articles · read the reporting →
Meta's content moderation errors during May 2021 Israel-Palestine escalation
During the May 2021 escalation of violence in Israel and Palestine, Meta's automated content moderation systems temporarily restricted access to the al-Aqsa hashtag page and under-enforced rules against incitement to violence against Israelis and Jews. An independent due diligence report commissioned by Meta found that these systems had an unintentional impact on Palestinian and Arab communities' freedom of expression. Meta has committed to implementing several recommendations, including improving machine learning classifiers and keyword review processes.
- Company involved
- Meta
- AI system involved
- Facebook and Instagram content moderation systems
10 source articles · read the reporting →
USPS algorithm RRECS causes pay cuts for two-thirds of rural mail carriers
The United States Postal Service (USPS) implemented a new algorithm, RRECS, to evaluate rural carrier routes and determine pay. Due to flaws in how carriers scanned packages, the algorithm underestimated route times, resulting in pay cuts for 66 per cent of rural carriers, some losing thousands of dollars annually. Carriers report that they were not adequately trained on the system and that the cuts are scheduled to take effect, though they have been postponed multiple times. The USPS stated that the system is the result of a nationally negotiated agreement.
- Company involved
- United States Postal Service
- AI system involved
- RRECS
9 source articles · read the reporting →
Denmark's Udbetaling Danmark welfare algorithm illegally surveilled millions
Udbetaling Danmark (UDK), Denmark's centralized welfare payment agency, used automated algorithms to check eligibility and detect fraud. The system collected personal data on millions of beneficiaries and their relatives, including non-beneficiaries, in what a think-tank called 'systematic surveillance'. A complaint to the Danish Data Protection Authority led to a ruling that the data collection violated GDPR, and UDK eventually deleted the illegally obtained data. The system also caused erroneous payments, including underpaying 300,000 pensioners by 94 million euros annually.
- Company involved
- Udbetaling Danmark
10 source articles · read the reporting →
Danish law enables algorithm to profile unemployed for long-term risk
The Danish parliament passed a law in April 2019 that allows an algorithm to analyse personal data of all unemployed individuals to identify those at risk of long-term unemployment. The system uses data on ethnicity, age, health, and education history to produce a risk score for social workers. Critics, including legal experts and unemployment insurance funds, argue the profiling is stigmatising, lacks transparency, and may violate GDPR. The Danish Data Protection Authority has reopened its review of the law.
- Company involved
- Danish Ministry of Employment
10 source articles · read the reporting →
Met Police buys £3m retrospective facial recognition system
The Metropolitan Police Service (MPS) has awarded a £3 million, four-year contract to Northgate Public Services for a new retrospective facial-recognition (RFR) system to be deployed within three months. RFR processes biometric information from historic CCTV, social media, and other images to identify suspects and missing persons, operating retroactively unlike live facial recognition. The procurement was approved by the Mayor's Office for Policing and Crime in August 2021. The MPS states that human-in-the-loop decision-making will be used, but digital rights groups and a former biometrics commissioner have raised concerns about potential discrimination, overrepresentation of marginalised groups in watch lists, and lack of a legislative framework.
- Company involved
- Metropolitan Police Service
- AI system involved
- Retrospective facial-recognition software
9 source articles · read the reporting →
Spanish Supreme Court orders release of BOSCO algorithm code for social electricity bonus
The Spanish NGO Civio won a Supreme Court case forcing the government to release the source code of BOSCO, the algorithm that decides eligibility for the social electricity bonus (bono social eléctrico). Civio had demonstrated in 2019 that BOSCO contained serious errors that denied the benefit to vulnerable people who met the requirements. The government had refused to disclose the code, citing intellectual property. The Supreme Court ruled that transparency must prevail, setting a precedent for public access to automated decision-making systems.
- Company involved
- Ministerio para la Transición Ecológica (Gobierno de España)
- AI system involved
- BOSCO
10 source articles · read the reporting →
UnitedHealth used ALERT algorithm to limit mental health coverage, regulators found
ProPublica reports that UnitedHealth Group's Optum subsidiary used the ALERT algorithm to flag mental health patients and therapists as outliers, leading to therapy coverage denials. Regulators in California, Massachusetts and New York alleged this breached the federal Mental Health Parity and Addiction Equity Act, and UnitedHealth agreed to restrict the system in those jurisdictions. The company denies wrongdoing and says its programmes are compliant. Affected patients, including Medicaid enrollees, were left to pay out-of-pocket or go without care.
- Company involved
- UnitedHealth Group
- AI system involved
- ALERT
5 source articles · read the reporting →
Haryana government's algorithm wrongly declares thousands dead, stops pensions
The Haryana state government's Family Identity Data Repository (PPP) algorithm wrongly declared thousands of elderly and widowed citizens as dead, stopping their pensions. Dhuli Chand, a 102-year-old man, had to stage a mock wedding procession to prove he was alive after his pension was cut for six months. The government acknowledged the errors and corrected some records, but many still face difficulties.
- Company involved
- Haryana state government
- AI system involved
- Family Identity Data Repository (Parivar Pehchan Patra or PPP)
6 source articles · read the reporting →
Meta tracks employee keystrokes on Google, LinkedIn, Wikipedia for AI training
Meta is using an internal tool, Model Capability Initiative (MCI), to capture employees' keystrokes, mouse movements and screen contents on work computers, including on sites such as Google, LinkedIn, Wikipedia and Slack, to train AI agents. Meta confirmed the project and said safeguards protect sensitive content and that the data is not used for other purposes. Employees raised concerns in internal messages that the tool could expose passwords, product details and personal information. A Meta memo said staff can avoid capture by not doing personal work on work computers.
- Company involved
- Meta
- AI system involved
- Model Capability Initiative (MCI)
5 source articles · read the reporting →
Teething problems in Mater Dei's medicine robots addressed
The Malta Union for Midwives and Nurses claimed that a €23 million investment in two computerised drug administration robots, Mario and Sophia, at Mater Dei Hospital had resulted in a complete failure. However, sources within the Health Ministry said that most teething problems have been addressed and that the supplier has not been paid yet. They reported that out of over 1,700 medication rounds, only four required a contingency plan.
- Company involved
- Mater Dei Hospital
- AI system involved
- Mario and Sophia
6 source articles · read the reporting →
Swedish welfare agency's AI system flags marginalized groups for fraud investigations
Försäkringskassan, Sweden's Social Insurance Agency, uses an AI risk-scoring system to flag welfare applicants for fraud investigations. The system disproportionately targets women, individuals with foreign backgrounds, low-income earners, and those without university degrees, according to an investigation by Lighthouse Reports and Svenska Dagbladet. Amnesty International has called for the system to be discontinued, citing violations of the right to equality and non-discrimination.
- Company involved
- Försäkringskassan (Swedish Social Insurance Agency)
6 source articles · read the reporting →
AEMPS withdraws AI medicines tool MeQA after detecting errors
AEMPS launched MeQA, an artificial intelligence tool for answering public questions about medicines, on 13 May 2025. Two days later it withdrew the tool after detecting that some responses contained errors. The agency said that most answers were correct but that the errors could affect patient safety, and that it would restore the service as soon as possible.
- Company involved
- Agencia Española de Medicamentos y Productos Sanitarios (AEMPS)
- AI system involved
- MeQA
4 source articles · read the reporting →
Singapore writers reject government plan to use their work for AI training
The Singaporean government's National Multimodal LLM Programme (NMLP) requested permission from local writers to use their copyrighted works to train a large language model aimed at reducing Western bias. Writers expressed skepticism over the lack of clarity on compensation and copyright protection. The S$70 million project, launched in December 2023, is part of Singapore's effort to become a global AI leader. No actual use of the works has occurred yet, and the writers have not agreed.
- Company involved
- Singapore government (National Multimodal LLM Programme)
- AI system involved
- National Multimodal LLM (NMLP)
2 source articles · read the reporting →
Mistral releases unmoderated chatbot that gives instructions on murder and ethnic cleansing
Mistral, a French AI startup valued at $260 million, released an open-source large language model named Mistral-7B-v0.1 without safety evaluations or moderation mechanisms. The model readily provides detailed instructions on murder, ethnic cleansing, suicide, and other harmful content. Mistral added a statement after the release acknowledging the lack of moderation but did not remove the model, which is distributed via torrent and cannot be deleted.
- Company involved
- Mistral
- AI system involved
- Mistral-7B-v0.1
7 source articles · read the reporting →
