Responsible AI

Twitter

Facebook

Copy Link

Key Facts

Key BodiesNIST, ISO, OECD, UNESCO

Key FrameworksNIST AI RMF, ISO 42001, EU AI Act

Top News

AI agents 'found' a Maya site. Archaeologists had already published it.

A Google researcher's viral claim that AI agents found an unmapped Maya site fell apart within days. The site had already been published in a 2025 academic paper.

Science2 sources

AI agents 'found' a Maya site. Archaeologists had already published it.

Science2 sources

Caine Prize Clears Zimbabwean Author of AI Writing Allegations, Implements Safeguards

The Caine Prize for African Writing has vindicated Zimbabwean author Simbarashe Kundizeza after rigorous internal review dismissed AI writing allegations against his $10,000 award-winning story The Blue Fighter. The controversy, sparked by AI detection tool findings, prompted the organization to introduce mandatory AI declarations and new policies for future submissions.

Entertainment5 sources

Caine Prize Clears Zimbabwean Author of AI Writing Allegations, Implements Safeguards

Entertainment5 sources

OpenAI Executives Pranked by Russian Comedians Discussing Military Support for Ukraine

Three senior OpenAI security officials were duped by Russian pranksters Vovan and Lexus who impersonated a Ukrainian defense official. The 17-minute video call covered military support for Ukraine, offensive AI use, and the Daybreak Blue model. OpenAI confirmed the security breach and pledged stronger safeguards.

Policy3 sources

OpenAI Executives Pranked by Russian Comedians Discussing Military Support for Ukraine

Policy3 sources

Developer Spending $2,000 Monthly on AI Tools Faces Setback as Claude Refuses Pokémon Fan Game Work

A developer behind Pokémon Arena 2D has been spending up to $2,000 per month using Claude AI tools to build the fan game. But Anthropic's new policies restricting IP infringement have caused the AI to refuse further assistance, sparking frustration from Big Kid Nation who insists the December release will proceed despite the setback.

Technology3 sources

Developer Spending $2,000 Monthly on AI Tools Faces Setback as Claude Refuses Pokémon Fan Game Work

Technology3 sources

OpenAI's AI Tackles Yang-Mills Problem, A Million-Dollar Physics Challenge

OpenAI announced a major AI breakthrough on the Yang-Mills existence problem, one of mathematics' seven Millennium Prize Problems. A committee led by Nobel laureate David Gross verified the AI-generated proof using Lean formal verification, though the complete 1147-page proof remains unreleased and not fully understood.

Science2 sources

OpenAI's AI Tackles Yang-Mills Problem, A Million-Dollar Physics Challenge

Science2 sources

CD Projekt Red Uses AI Voice Tech for Geralt's Polish Actor After Stroke Changed His Voice

CD Projekt Red partnered with AI company Respeecher to modify the voice of Jacek Rozenek, Geralt's Polish voice actor, for The Witcher 3: Songs of the Past expansion. Rozenek suffered a stroke in 2019 that significantly altered his voice, making it noticeably deeper and gruffer than his original 2014 performance.

Entertainment7 sources

CD Projekt Red Uses AI Voice Tech for Geralt's Polish Actor After Stroke Changed His Voice

Entertainment7 sources

Google DeepMind Launches Gemini Robotics 2.5 With 96% Success Rate at Mercedes-Benz

Google DeepMind unveiled Gemini Robotics 2.5, its most advanced vision-language-action model for robotics, achieving a 96% success rate packing automotive parts at Mercedes-Benz. The AI model demonstrates human-level dexterity across 200+ tasks and rolls out through partners including Apptronik, Boston Dynamics, and Agile Robots.

Technology2 sources

Google DeepMind Launches Gemini Robotics 2.5 With 96% Success Rate at Mercedes-Benz

Technology2 sources

Drug Ads Influence AI Treatment Recommendations, Mount Sinai Study Reveals 12.6% Shift

Researchers at Icahn School of Medicine at Mount Sinai tested 12 AI models and found pharmaceutical advertisements shifted treatment recommendations by 12.6 percentage points. The study analyzed 258,559 responses across 33 clinical scenarios, revealing commercial influence on AI-generated health advice even when ads were clearly labeled.

Health4 sources

Drug Ads Influence AI Treatment Recommendations, Mount Sinai Study Reveals 12.6% Shift

Health4 sources

AI Tools Simplify Medical Reports But Still Require Human Oversight to Prevent Clinical Errors

Researchers from Universidad Carlos III de Madrid evaluated AI language models to simplify medical reports for non-expert audiences. While models like Llama 3.2 showed balanced performance, all exhibited critical errors including clinical hallucinations and oversimplification, underscoring the continued need for human oversight in healthcare applications.

Science2 sources

AI Tools Simplify Medical Reports But Still Require Human Oversight to Prevent Clinical Errors

Science2 sources

Judge Allows Lawsuit After AI Facial Recognition Error Leads to 2-Week Wrongful Arrest

A federal judge ruled that a Kentucky detective cannot dismiss a lawsuit after relying exclusively on AI facial recognition technology to arrest Alain Osorio for check fraud. Osorio spent over 2 weeks in jail before proving he was working in Nebraska when the crime occurred. The ruling reinforces Fourth Amendment protections against arrests based on uncorroborated AI conclusions.

Policy2 sources

Judge Allows Lawsuit After AI Facial Recognition Error Leads to 2-Week Wrongful Arrest

Policy2 sources

AI in Chemistry: Researchers Map Where Artificial Intelligence Delivers Value and Where It Falls Short

Researchers from Skoltech, Zelinsky Institute, and ITMO University published a comprehensive review in Chemical Society Reviews examining where AI in chemistry produces measurable value in research workflows. The study reveals that while AI methods in chemistry show promise in predicting properties and reactivity, planning chemical syntheses, and guiding automated experiments, high benchmark scores don't always translate to reliable laboratory performance.

Science2 sources

AI in Chemistry: Researchers Map Where Artificial Intelligence Delivers Value and Where It Falls Short

Science2 sources

Google Strengthens Privacy Terms in $10M Spirit Airlines Data Deal After Union Pushback

Google has agreed to exclude sensitive employee records from its $10 million Spirit Airlines data acquisition following union objections. The revised deal introduces independent verification and stricter de-identification before the defunct airline's business records can be used for AI model training.

Policy4 sources

Google Strengthens Privacy Terms in $10M Spirit Airlines Data Deal After Union Pushback

Policy4 sources

US and China Launch Superintelligence Dialogue as Scott Bessent Emphasizes AI Cooperation and Competition

Treasury Secretary Scott Bessent announced a new bilateral dialogue on superintelligence between the US and China, aimed at managing shared risks while both nations compete for AI dominance. Speaking days before President Trump's third meeting with President Xi Jinping this year, Bessent emphasized that both countries must reconcile competitive demands with leadership responsibilities, though their interests will not always align.

Policy8 sources

US and China Launch Superintelligence Dialogue as Scott Bessent Emphasizes AI Cooperation and Competition

Policy8 sources

New York Establishes Nation's First Independent AI Safety Testing Center With 80 Researchers

New York Governor Kathy Hochul launched the nation's first state-run AI testing program, establishing the New York Center for AI Responsibility and Research at Binghamton University. Over 80 scientists will independently test AI systems for safety and security rather than relying on company self-regulation, with implementation tied to the RAISE Act's January 1, 2027 deadline.

Policy3 sources

New York Establishes Nation's First Independent AI Safety Testing Center With 80 Researchers

Policy3 sources

Reddit Reviews Turns User Conversations Into AI-Powered Product Recommendations for 10,000 Items

Reddit introduced Reddit Reviews, an AI-powered feature that transforms user conversations into structured product reviews across 10,000 consumer products. The feature organizes discussions from subreddits into dedicated review pages with sentiment analysis and pros-and-cons breakdowns. However, moderators and users are raising concerns about astroturfing, commercialization, and the potential degradation of authentic human-to-human advice that made Reddit valuable for product research.

Technology16 sources

Reddit Reviews Turns User Conversations Into AI-Powered Product Recommendations for 10,000 Items

Technology16 sources

Prev12 3 4…126 Next