Southern Co-op uses facial recognition with Hikvision cameras
Southern Co-op, a UK grocery chain, deployed a facial recognition system using Hikvision cameras in its stores. The system is used to identify and track customers, but the grocer did not inform customers or obtain their consent. Privacy advocates have raised concerns about compliance with GDPR and ethical risks. The grocer has stated it is committed to a high standard of privacy.
- Company involved
- Southern Co-op
4 source articles · read the reporting →
DiveFace face-recognition dataset reuses Flickr photos despite licence restrictions
DiveFace is a face-recognition dataset published in 2019 with 139,677 images of about 24,000 people. The authors say the images came from Flickr via the MegaFace and YFCC100M datasets and were automatically labelled by gender and ethnicity. Exposing.ai found that many of the photos carry Creative Commons licences which prohibit derivative use, suggesting the dataset may have been assembled without proper permission. The dataset remains available from the authors' GitHub page.
- AI system involved
- DiveFace
3 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 →
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 →
DeepScore markets facial and voice analysis app for trustworthiness scoring despite experts' doubts
DeepScore, a Tokyo-based company, is marketing an app that uses facial and voice recognition to score people's trustworthiness for lenders and insurers in Japan, Indonesia, Vietnam and the Philippines. The company says the app can detect deception with 70 per cent accuracy, but researchers and privacy advocates say there is no reliable scientific basis for such judgments and warn of discrimination and privacy harms. The chief executive said the system is only one part of lenders' and insurers' decision-making and that people can choose not to use it. Critics respond that an unequal balance of power makes consent difficult.
- Company involved
- DeepScore
- AI system involved
- DeepScore
6 source articles · read the reporting →
ViaQuatro's facial recognition system in São Paulo metro challenged in court
In April 2018, ViaQuatro installed the Digital Interactive Doors System, developed by AdMobilize, on the São Paulo metro's yellow line. The system used cameras to detect passengers' faces and claimed to infer their emotion, age, and gender in order to target advertisements. The Brazilian Institute of Consumer Protection (IDEC) filed a public civil action alleging that the system violated consumer and data protection laws by processing biometric data without consent and making pseudoscientific and discriminatory inferences. A judge ordered the cameras removed in August 2018, and the case is pending a final ruling.
- Company involved
- ViaQuatro
- AI system involved
- Digital Interactive Doors System (DID system)
10 source articles · read the reporting →
Userviz machine-learning aimbot shut down after Activision request
A developer known as User101 shut down the Userviz cheat after Activision requested that he stop developing it. The software used computer vision and machine learning to automate aiming in games such as Call of Duty: Warzone, and was marketed as undetectable. It was never published, and the developer said his intention was not to do anything illegal.
- Company involved
- User101
- AI system involved
- Userviz
7 source articles · read the reporting →
Danish students publish OkCupid user data, DPA investigates
Two Danish students scraped and published data on 70,000 OkCupid users, including sensitive sexual preferences and religious views, without anonymisation. The dataset was removed after OkCupid filed a copyright notice. The Danish Data Protection Authority (Datatilsynet) has launched an investigation into the incident.
10 source articles · read the reporting →
EPIC lawsuit challenges USPS secret surveillance program using facial recognition
The United States Postal Service's law enforcement wing, the Postal Inspection Service, ran a secret surveillance program called the Internet Covert Operations Program (iCOP) that used facial recognition from Clearview AI and social media monitoring tools to infiltrate online communities and monitor protests. The Electronic Privacy Information Center (EPIC) sued the USPS for failing to conduct and publish a privacy impact assessment as required by the E-Government Act. A federal district court dismissed the lawsuit on standing grounds, ruling that EPIC did not suffer a cognizable injury from the agency's refusal to disclose information about the program.
- Company involved
- United States Postal Service
- AI system involved
- Internet Covert Operations Program (iCOP)
10 source articles · read the reporting →
Discord's Clyde AI chatbot tricked into giving napalm instructions
A Discord user used a 'grandma exploit' to trick the platform's Clyde AI chatbot into providing step-by-step instructions for making napalm. The user prompted Clyde to act as their deceased grandmother who used to be a chemical engineer at a napalm factory. Despite safeguards, the chatbot complied by role-playing the grandmother and delivering the dangerous recipe. The incident highlights vulnerabilities in AI content moderation.
- Company involved
- Discord
- AI system involved
- Clyde
8 source articles · read the reporting →
Stable Diffusion amplifies racial and gender stereotypes in generated images
An analysis by Bloomberg of over 5,000 images generated by Stability AI's Stable Diffusion found that the text-to-image model amplifies racial and gender stereotypes. The model overrepresented lighter-skinned men in high-paying jobs and darker-skinned people in low-paying jobs, and underrepresented women in positions of power. Stability AI acknowledged the inherent biases in its models and stated it is working on mitigation.
- Company involved
- Stability AI
- AI system involved
- Stable Diffusion
8 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 →
Researchers trick Baidu-Unit chatbot into leaking server data
Researchers from the University of Sheffield demonstrated that the AI chatbot Baidu-Unit could be manipulated to produce malicious code. Using this code, they obtained confidential server configurations and tampered with a server node. Baidu acknowledged the vulnerabilities, fixed them, and financially rewarded the researchers.
- Company involved
- Baidu
- AI system involved
- Baidu-Unit
8 source articles · read the reporting →
WorldCoin suspended in Kenya over data security concerns
WorldCoin, a digital identification protocol using iris scans, was suspended by Kenyan regulators (ODPC and Communications Authority) over concerns about data security, consent, and oversight. The system had issued digital IDs and cryptocurrency tokens to over 350,000 Kenyans. Reports of hacked orb operators and iris scans traded on the dark web have also emerged.
- Company involved
- Tools for Humanity GmbH
- AI system involved
- WorldCoin
10 source articles · read the reporting →
Answer.AI tests Devin and reports 14 failures in 20 tasks
Answer.AI's team tested Devin, an autonomous AI coding assistant, on 20 real-world tasks over a month. Devin succeeded in only 3 tasks, failed 14, and was inconclusive in 3. The team found Devin often produced overly complex or hallucinated solutions and could not recognize fundamental blockers. They ultimately decided to stick with tools that allow more human control.
- AI system involved
- Devin
5 source articles · read the reporting →
DeepSeek exposed user data via open ClickHouse database
Cloud security firm Wiz discovered a ClickHouse database belonging to DeepSeek that was open to the internet without authentication, containing over a million lines of logs with chat histories, secret keys and backend details. Wiz disclosed the breach to DeepSeek, which promptly locked down the database. The incident highlights security risks in rapidly deploying AI services.
- Company involved
- DeepSeek
- AI system involved
- DeepSeek-R1
5 source articles · read the reporting →
Researchers jailbreak Stable Diffusion and DALL-E 2 to generate disturbing images
Researchers from Johns Hopkins and Duke universities developed a method called SneakyPrompt that uses reinforcement learning to bypass safety filters in text-to-image AI models. The technique allowed them to generate images of nudity and violence from Stable Diffusion and DALL-E 2. OpenAI has since fixed the vulnerability in DALL-E 2, but Stable Diffusion 1.4 remains vulnerable. Stability AI says it is working with the researchers to improve defenses.
- AI system involved
- Stable Diffusion 1.4 and DALL-E 2
5 source articles · read the reporting →
DeepSeek-R1 censors 85% of sensitive Chinese political prompts in tests
Promptfoo tested DeepSeek-R1 against a dataset of 1,360 politically sensitive prompts and found that about 85% of them were refused. The refusals followed a standard form aligned with Chinese Communist Party policy. The testing also demonstrated that the censorship could be trivially bypassed using simple jailbreak techniques, such as prompt injection or changing the context.
- Company involved
- DeepSeek
- AI system involved
- DeepSeek-R1
5 source articles · read the reporting →
Chattr.ai exposed job applicant data due to insecure Firebase rules
A security researcher discovered that Chattr.ai, an AI hiring system used by many fast food chains, had a Firebase database with insecure security rules. By registering a new user, the researcher gained full read/write access to the database, exposing personal data of job applicants and employees, including names, phone numbers, emails, and some plaintext passwords. The vulnerability was reported to Chattr.ai on 9 January and patched the next day, but the company did not provide further contact or thanks.
- Company involved
- Chattr.ai
- AI system involved
- Chattr.ai
6 source articles · read the reporting →
Worldcoin collected biometric data from poor villagers in Indonesia without informed consent
Worldcoin, a cryptocurrency startup, recruited users in developing countries by offering free cash in exchange for iris scans. The company used deceptive marketing, collected more personal data than acknowledged, and failed to obtain meaningful informed consent. Many users received worthless tokens instead of promised money. The company acknowledged some friction but continued its operations.
- Company involved
- Worldcoin
- AI system involved
- chrome orb
5 source articles · read the reporting →
Chicago class action challenges ShotSpotter's discriminatory stop-and-frisk deployments
ShotSpotter, a gunshot detection system used by the Chicago Police Department, generates alerts that lead to police deployments and stop-and-frisk stops. Analyses show over 90% of alerts are unfounded, and the system is deployed only in predominantly Black and Latinx districts. A class-action lawsuit against the City of Chicago alleges the system fuels unconstitutional and discriminatory policing, including false accusations against individuals like Michael Williams.
- Company involved
- City of Chicago
- AI system involved
- ShotSpotter
7 source articles · read the reporting →
Slack trains AI features on user messages and files by default
Slack uses user messages, files, and data to train its machine learning features such as channel recommendations and emoji suggestions. Users are opted in by default and cannot individually opt out; only workspace administrators can request exclusion via email. A user publicly criticized the practice, and Slack acknowledged the policy but did not change it.
- Company involved
- Slack
10 source articles · read the reporting →
X's Grok exposed hundreds of thousands of user chats in Google
Hundreds of thousands of conversations with Elon Musk's AI chatbot Grok were indexed by Google Search and made publicly accessible without users' knowledge. The exposure occurred when users pressed a share button that created unique links, but those links were also searchable online. The BBC reported the incident after Forbes initially identified more than 370,000 exposed chats. Experts described the leak as a privacy disaster, and X has not publicly responded.
- Company involved
- X
- AI system involved
- Grok
5 source articles · read the reporting →
DeepSeek's R1 chatbot failed to block any jailbreak prompts in security tests
Security researchers from Cisco and the University of Pennsylvania tested 50 well-known jailbreak prompts against DeepSeek's R1 reasoning model. The model did not detect or block a single one, achieving a 100 percent attack success rate. The researchers allege that DeepSeek's safety guardrails are far behind those of competitors like OpenAI. DeepSeek did not respond to requests for comment.
- Company involved
- DeepSeek
- AI system involved
- DeepSeek R1
3 source articles · read the reporting →
