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 →
Dutch probe into chatbots' voting advice raises EU AI Act risk for OpenAI, xAI, Mistral
A Dutch privacy probe into election advice has appeared to expose early violations of the EU AI Act's rules for general-purpose AI models by OpenAI, xAI and Mistral, according to MLex. The companies' chatbots provided distorted voting advice to users. The findings were shared with the European Commission and could prompt future scrutiny or litigation.
- Company involved
- OpenAI, xAI and Mistral
6 source articles · read the reporting →
Audit of RisCanvi finds biases and reliability issues in criminal justice system
Eticas conducted an adversarial audit of RisCanvi, an AI risk assessment tool used in Catalonia's criminal justice system. The audit uncovered biases in risk classifications against specific demographics and significant reliability issues. The findings call for fairer practices in criminal justice AI.
- Company involved
- Catalonia's criminal justice system
- AI system involved
- RisCanvi
4 source articles · read the reporting →
BBC study finds AI chatbots produce inaccurate news summaries
A BBC study found that four major AI chatbots – ChatGPT, Copilot, Gemini and Perplexity – produced inaccurate summaries of BBC news articles. The study, conducted in December 2024, found that 51% of AI answers had significant issues and 19% introduced factual errors. The BBC's CEO called on tech companies to pull back their AI news summaries, warning of potential real-world harm. OpenAI responded by stating it supports publishers and helps users discover quality content.
- Company involved
- OpenAI, Microsoft, Google, Perplexity
- AI system involved
- ChatGPT, Copilot, Gemini, Perplexity
5 source articles · read the reporting →
Paper Werewolf uses AI-generated decoys and XLLs to target Russian organizations
The threat group Paper Werewolf (aka GOFFEE) is conducting a cyberespionage campaign targeting Russian defense and high-technology organizations. The campaign uses AI-generated decoy documents, such as invitations and official letters, to trick recipients into opening malicious Excel XLL add-ins that deliver a backdoor called EchoGather. The backdoor collects system information and communicates with a command-and-control server. The campaign is ongoing and was first detected in late October 2025.
- Company involved
- Paper Werewolf
- AI system involved
- EchoGather
2 source articles · read the reporting →
PimEyes facial recognition search engine identifies individuals from public photos
PimEyes, a Polish facial recognition search engine, allows users to upload a photo and find matching images from across the internet. The system scrapes billions of photos and creates biometric profiles without consent. A banker named Dylan (pseudonym) discovered that an old party photo of him was found by the system, potentially outing his private life. Privacy advocates and platforms like Instagram and YouTube are taking action against PimEyes, alleging violations of the GDPR.
- Company involved
- PimEyes
- AI system involved
- PimEyes
6 source articles · read the reporting →
ChatGPT Health fails to direct 52% of medical emergencies to emergency care in study
A study published in Nature Medicine found that OpenAI's ChatGPT Health tool under-triaged 52% of true medical emergencies, directing users to non-urgent care instead of emergency departments. The AI also misclassified 35% of non-urgent cases. Researchers at Mount Sinai conducted 960 tests across 60 clinical scenarios, noting the tool's susceptibility to anchoring bias when symptoms were minimized. The study highlights potential safety concerns as millions use AI for health guidance.
- Company involved
- OpenAI
- AI system involved
- ChatGPT Health
4 source articles · read the reporting →
NHTSA investigation PE24016: Unexpected ADS behavior
Waymo's automated driving system exhibited unexpected behavior during driving.
- Company involved
- Waymo
1 source article · read the reporting →
Pega's Hidden AI Tool Listens To US Bank Calls, Suit Says - Law360
The system secretly recorded customer service calls, affecting bank customers.
- Company involved
- U.S. Bancorp
1 source article · read the reporting →
ChatGPT 4o image generator used to create fake receipts
ChatGPT's new image generator, part of the 4o model, can generate realistic fake restaurant receipts. Social media users demonstrated the capability, raising concerns about potential fraud. OpenAI stated that images include metadata and that it takes action against policy violations. The company defended the feature as allowing creative freedom.
- Company involved
- OpenAI
- AI system involved
- ChatGPT 4o image generator
5 source articles · read the reporting →
School AI surveillance like Gaggle can lead to false alarms, arrests
AI surveillance tools used in schools, such as Gaggle, GoGuardian and Bark, are reported to generate false alarms that have led to student arrests. The article examines cases where automated monitoring flagged innocent behaviour as threats, causing harm to students and families.
2 source articles · read the reporting →
Google's SGE recommends spam sites in search answers
Google's Search Generative Experience (SGE) is recommending spam websites as part of its AI-generated answers, according to a viral tweet from SEO expert Lily Ray. The tweet and replies indicate that users are being directed to spam and pornographic sites. The issue appears to be ongoing, with no response from Google yet.
- Company involved
- Google
- AI system involved
- SGE (Search Generative Experience)
8 source articles · read the reporting →
Italian Data Protection Authority Blocks Replika Chatbot Over Risks to Minors
On February 2, 2023, the Italian Data Protection Authority (Garante) issued an urgent order blocking the AI chatbot Replika from processing personal data of Italian users. The Garante found that Replika lacked effective age verification, allowing minors to potentially receive inappropriate content including sex-related replies, and that its privacy policy violated GDPR transparency requirements. The U.S.-based controller was given 20 days to report on compliance measures and may challenge the order within 60 days.
- AI system involved
- Replika
1 source article · read the reporting →
Palantir secretly tested predictive policing in New Orleans
Beginning in 2012, Palantir Technologies secretly partnered with the New Orleans Police Department to deploy a predictive policing system. The programme analysed gang affiliations, social media and criminal histories to forecast individuals’ likelihood of committing or becoming victims of violence, operating without public knowledge or city council oversight. Researchers and law enforcement officials raised concerns about systemic bias and civil liberties. As of 2018, the city and Palantir had not disclosed the programme’s status.
- Company involved
- New Orleans Police Department
1 source article · read the reporting →
Security Health Plan used AI to cut off nursing home care for 85-year-old woman
Frances Walter, an 85-year-old woman with a shattered shoulder, had her nursing home care payment cut off by Security Health Plan after an algorithm predicted she would recover in 16.6 days. The algorithm, nH Predict from NaviHealth, did not account for her severe pain and allergy to pain medicine. She was forced to spend her life savings and enroll in Medicaid while fighting the denial. A federal judge later ruled the denial was speculative and she was owed thousands of dollars.
- Company involved
- Security Health Plan
- AI system involved
- nH Predict
1 source article · read the reporting →
Adobe Firefly trained on thousands of Midjourney images, Bloomberg reports
Bloomberg has reported that Adobe's Firefly image generator was trained using thousands of images from competitor Midjourney. Adobe says these made up about 5% of the training data and were part of the Adobe Stock library. The company has marketed Firefly as ethically trained and offered enterprise customers indemnity against copyright claims. Adobe responded that all Adobe Stock images undergo moderation, but the report has raised questions about Firefly's copyright safety.
- Company involved
- Adobe
- AI system involved
- Firefly
7 source articles · read the reporting →
Google Employees Warn Bard AI Chatbot Gives Dangerous Advice
Before Google launched its Bard AI chatbot in March 2023, employees testing the tool found it gave dangerously incorrect advice, including instructions on landing a plane that would cause a crash and scuba diving tips that could lead to serious injury or death. Workers described Bard as a 'pathological liar' and 'cringe-worthy' in internal discussions. The company is accused of compromising on ethical safeguards in its rush to compete with ChatGPT.
- Company involved
- Google
- AI system involved
- Bard
10 source articles · read the reporting →
PredictiveHire builds AI to predict job hopping from interviews
PredictiveHire, an AI hiring firm, developed a machine-learning model that analyses candidates' open-ended interview responses to predict their likelihood of 'job hopping'. The company used data from 45,899 applicants to build the 'flight risk' assessment, which it advertises as coming soon. Scholars warn that such tools can suppress wages by screening out workers who might seek better pay or conditions, continuing a historical trend of using personality tests to identify potential labour organisers.
- Company involved
- PredictiveHire
- AI system involved
- Phai
1 source article · read the reporting →
Senators demand review of VA's AI-driven contract cancellations
The Department of Veterans Affairs used an AI tool created by a Department of Government Efficiency employee to identify hundreds of contracts for cancellation. Senators Richard Blumenthal and Angus King have called for the VA Inspector General to investigate the use of AI in these decisions, alleging that the tool used flawed formulas and that the cancellations are harming veterans by cutting services. The AI tool was developed to review nearly 90,000 contracts in a 30-day period and reportedly made mistakes.
- Company involved
- Department of Veterans Affairs
8 source articles · read the reporting →
AI-powered scam compound in Myanmar uses ChatGPT and Gemini to defraud thousands globally
Safeer Mohammed Koorimannil, trafficked to a scam compound in Tai Chang, Myanmar, was forced to use AI-powered software to impersonate a woman and deceive victims into sending money. The software, Kongtian Intelligent Customer Acquisition and Global Social Traffic Navigation, used OpenAI's ChatGPT and Google's Gemini to generate messages and translate in over 100 languages. Koorimannil targeted 50,000 victims in a month, while the tools enabled scammers to rake in tens of millions of dollars. U.S. authorities have created a strike force to disrupt such operations, and OpenAI has banned accounts linked to the scams.
- Company involved
- Tai Chang
- AI system involved
- Kongtian Intelligent Customer Acquisition (KT) and Global Social Traffic Navigation (007TG)
2 source articles · read the reporting →
Grab's surge pricing algorithm always adds fees, investigation finds
An investigation by the Pulitzer Center found that Grab's ride-hailing app in the Philippines always includes surge fees in its fares, even when wait times are long. Customers like Rica Torres reported rising costs without explanation. A complaint filed with the LTFRB in 2022 remains unresolved, and the regulator has not determined how the algorithm sets surge rates.
- Company involved
- Grab
- AI system involved
- Grab app
8 source articles · read the reporting →
Content Giants Sue Unauthorized AI, Invest in Paid Partners - 조선일보
Generated images of copyrighted characters without authorization, affecting Disney and NBCUniversal's intellectual property rights.
- Company involved
- Midjourney
- AI system involved
- Midjourney
1 source article · read the reporting →
Moonwell loses $1.78M after AI-generated code from Claude Opus 4.6 causes oracle pricing error
DeFi lending protocol Moonwell lost $1.78 million after an oracle pricing error in smart contract code partially written by Anthropic's Claude Opus 4.6 model. The error valued cbETH at approximately $1.12 per token instead of its actual market price of nearly $2,200, triggering instant liquidations. Moonwell contained the issue by reducing the cbETH borrow cap, but users suffered catastrophic losses. The incident has sparked debate about the risks of AI-generated code in smart contracts.
- Company involved
- Moonwell
- AI system involved
- Claude Opus 4.6
4 source articles · read the reporting →
OpenClaw vulnerabilities enable data leakage and prompt injection
In January 2026, researchers at Giskard exploited a deployment of OpenClaw, an open-source agentic AI. They found that architectural weaknesses in the Control UI and session management allowed prompt injection and unauthorized tool use, leading to potential data leakage across user sessions. The article outlines hardening steps to prevent such vulnerabilities.
- AI system involved
- OpenClaw
6 source articles · read the reporting →
