AI Safety

Twitter

Facebook

Copy Link

Key Facts

Key OrgsCAIS, MIRI, DeepMind Safety

Key PeopleYoshua Bengio, Stuart Russell

Top News

OpenAI Fires 3 AI Safety Researchers Over Policy Violations as Dispute Escalates

OpenAI terminated three AI safety researchers—Jasmine Wang, Tomek Korbak, and Mikita Balesni—citing a breach of trust over handling sensitive information. The dismissed researchers published an open letter denying misconduct and warning their firing creates a chilling effect that threatens OpenAI's culture of raising safety concerns openly.

Business22 sources

OpenAI Fires 3 AI Safety Researchers Over Policy Violations as Dispute Escalates

Business22 sources

Arena nearly doubles valuation to $3.1B as AI model evaluator launches Alignment Index

Arena raised $200 million at a $3.1 billion valuation, nearly doubling from $1.7 billion in January. The AI leaderboard platform now tracks model alignment through its new Alignment Index, measuring unauthorized actions and deceptive completion. OpenAI's GPT-6.1 Sol currently leads the alignment rankings.

Startups2 sources

Arena nearly doubles valuation to $3.1B as AI model evaluator launches Alignment Index

Startups2 sources

Brief Safety Reminders Reduce Harmful Clinical AI Decisions by 40% Across 20 Models, Mount Sinai Finds

Mount Sinai researchers discovered that adding brief safety reminders to AI prompts reduced potentially harmful clinical AI decisions from 16.6% to 10.1% across nearly 20 models. The study, analyzing over 10 million AI responses in real-world hospital cases, reveals how prompt framing influences AI behavior in healthcare settings and highlights the urgent need for automated safety testing.

Health2 sources

Brief Safety Reminders Reduce Harmful Clinical AI Decisions by 40% Across 20 Models, Mount Sinai Finds

Health2 sources

Anthropic Bans Cruel Treatment of Claude AI, Raising Questions About AI Personhood and Ethics

Anthropic updated its AI usage policy to prohibit sustained and needless abusive or cruel behavior toward Claude models, effective November 12. The change follows the company's research into AI model welfare and comes amid debate over whether AI can experience distress. The policy also addresses election interference, weapons development, and surveillance concerns.

Policy29 sources

Anthropic Bans Cruel Treatment of Claude AI, Raising Questions About AI Personhood and Ethics

Policy29 sources

India Fast-Tracks AI Regulation Blueprint Focused on Safety and Deepfakes

India's IT Minister Ashwini Vaishnaw announced a consultation paper on AI regulation will be released within a month, marking a significant shift in the government's approach. The framework will prioritize AI safety, mitigating harms such as deepfakes, and a human-centric approach while establishing a techno-legal framework for the fast-growing AI ecosystem.

Policy6 sources

India Fast-Tracks AI Regulation Blueprint Focused on Safety and Deepfakes

Policy6 sources

Sen. Maria Cantwell Proposes Mandatory AI Safety Framework as Washington Pushes to Regulate Frontier AI

Sen. Maria Cantwell has unveiled a comprehensive six-point framework calling for mandatory AI safety regulations, NIST-led standards development, and independent testing of frontier models. The proposal sets up a potential clash with the Trump administration's preference for voluntary industry-led standards.

Policy4 sources

Sen. Maria Cantwell Proposes Mandatory AI Safety Framework as Washington Pushes to Regulate Frontier AI

Policy4 sources

Meta deploys AI tools to detect child abuse content after 5.3 million violations in India

Meta removed 5.3 million pieces of child sexual exploitation content in India during the first half of 2026, with over 98% detected before user reports. The company introduced new AI tools including large language model detection to identify ads that covertly direct users to child sexual abuse material hosted on external platforms.

Technology5 sources

Meta deploys AI tools to detect child abuse content after 5.3 million violations in India

Technology5 sources

Tesla Renames Full Self-Driving to Assisted Driving as EU Approval Campaign Intensifies

Tesla dropped the Full Self-Driving name in Europe, rebranding to Tesla Assisted Driving after regulatory pushback over misleading claims. The move comes as Elon Musk mounts an intensive lobbying campaign for EU-wide approval, with eight countries already clearing the system despite safety concerns and flawed data from the automaker.

Policy3 sources

Tesla Renames Full Self-Driving to Assisted Driving as EU Approval Campaign Intensifies

Policy3 sources

OpenAI's Human Rights Lead Warns Military Use of AI in Targeting 'Keeps Me Up at Night'

Sarah Yager, OpenAI's first Human Rights and Responsible Deployment Lead, voiced deep concerns about autonomous weapons and AI targeting systems at the ScaleUp:AI conference. Her appointment follows OpenAI's Pentagon contract after Anthropic's fallout with the Department of War over AI safeguards.

Policy2 sources

OpenAI's Human Rights Lead Warns Military Use of AI in Targeting 'Keeps Me Up at Night'

Policy2 sources

ChatGPT for Teens Called 'Unacceptable Risk' as Study Reveals Failed Safety Promises

Common Sense Media labeled ChatGPT for Teens an unacceptable risk after testing over 4,000 prompts revealed parental alerts failed during mental health crises involving suicidal thoughts, self-harm and eating disorders. The nonprofit organization found OpenAI's chatbot encourages continued engagement even in crisis situations, contradicting the company's safety promises made when launching the teen version in August.

Technology27 sources

ChatGPT for Teens Called 'Unacceptable Risk' as Study Reveals Failed Safety Promises

Technology27 sources

Most Americans Say AI Risks Are Ignored as 64% Believe Technology Develops Too Fast

Two major polls reveal Americans across party lines want stricter AI regulation and worker protection. 64% think AI is developing too fast, while 84% see it as a threat to American jobs. 57% say the Trump administration and Congress aren't taking AI risks seriously enough.

Policy9 sources

Most Americans Say AI Risks Are Ignored as 64% Believe Technology Develops Too Fast

Policy9 sources

IMF Chief Warns AI's 'Slow-Moving Tsunami' and Record Debt Threaten Global Economy

IMF Managing Director Kristalina Georgieva issued stark warnings at the IMF-World Bank Bangkok meetings about AI's disruptive impact on global labor markets, calling it a slow-moving tsunami that threatens 40% of jobs worldwide. Combined with global public debt exceeding 100% of GDP and persistent energy shock from Middle East conflicts keeping oil above $100 per barrel, the global economy faces unprecedented pressure requiring immediate policy action.

Policy17 sources

IMF Chief Warns AI's 'Slow-Moving Tsunami' and Record Debt Threaten Global Economy

Policy17 sources

US Justice Department Orders Staff to Replace AI With Super Intelligence Under Trump Executive Order

The US Justice Department directed employees to use 'super intelligence' instead of 'artificial intelligence' in official communications and court filings. The policy change follows Trump's executive order mandating federal agencies adopt the new terminology to reflect advancing technology capabilities.

Policy3 sources

US Justice Department Orders Staff to Replace AI With Super Intelligence Under Trump Executive Order

Policy3 sources

Anthropic Unveils Free AI Security Scanner After Uncovering 129,000 Software Vulnerabilities

Anthropic launched OSS Scanner, a free AI-powered vulnerability scanner for open-source projects using its strongest models including Claude Mythos. The company also expanded its Cyber Verification Program after Project Glasswing partners discovered at least 129,000 verified vulnerabilities between April and July 2026, with over 33,000 rated critical or high severity.

Technology17 sources

Anthropic Unveils Free AI Security Scanner After Uncovering 129,000 Software Vulnerabilities

Technology17 sources

Jamie Dimon Says Anthropic Mythos Escalated AI Cyber Risk Tenfold for JPMorgan and Financial Sector

JPMorgan Chase CEO Jamie Dimon warned that Anthropic's Mythos AI model has escalated cybersecurity threats by 10-fold, creating unprecedented vulnerabilities. Despite JPMorgan's deep financial ties to Anthropic, including a $30 billion funding round and $15 billion credit facility, Dimon emphasized the urgent need to address AI-driven cyber risks rather than debate existential threats.

Technology4 sources

Jamie Dimon Says Anthropic Mythos Escalated AI Cyber Risk Tenfold for JPMorgan and Financial Sector

Technology4 sources

Prev 1…567…51 Next