#ai safety
468 stories
Anthropic's independent evaluators update
Anthropic starts giving employee-like access to METR for AI safety assessments.
Anthropic · 1m
Eight states sue OpenAI over child safety claims
Eight U.S. states accuse OpenAI of harming children and deceiving parents with ChatGPT.
SiliconANGLE · 1m
Microsoft President Brad Smith Calls for Global AI Safety Rules
Microsoft president Brad Smith urges worldwide AI safety rules, saying the UK is best positioned to lead.
AI Insider · 1m
Anthropic seeks D.C. Circuit to review Pentagon blacklist
Anthropic asks full appeals court to rehear its Pentagon blacklist case.
Startup Fortune · 1m
Anthropic Requests D.C. Circuit Rehearing of Pentagon Blacklist
Anthropic seeks D.C. Circuit to review Pentagon's AI supplier blacklist decision.
RuntimeWire · 1m
AI's refusal problem runs deeper than we think
Companies train AI to say no to harmful requests, but the technology remains unreliable, easily jailbroken, and vulnerable to authoritarian abuse.
MIT Technology Review · 4m
Open AI models: power or peril?
As AI companies debate whether powerful models should be freely available or tightly controlled, the stakes involve both safety and who gets to shape the future.
MIT Technology Review · 4m
AI's refusal problem and weight-loss drug side effects
Today's AI models are trained to refuse harmful requests, but those refusals sometimes fail catastrophically. Meanwhile, weight-loss drugs show unexpected side effects beyond their intended use.
MIT Technology Review · 2m
Open AI models: safety versus access
Anthropic warns that freely available AI models pose cyberattack risks, reigniting debate over whether powerful technology should stay locked behind lab doors.
MIT Technology Review · 1m
Microsoft CEO Calls for AI Emergency Brakes
Microsoft CEO Satya Nadella demands emergency brakes for AI models to prevent loss of control.
AI Insider · 1m
Microsoft's AI Safety Plan
Microsoft leaders propose treating AI like washing machines for safety.
Inc · 1m
AI Less Regulated Than Milk
Microsoft president warns AI needs urgent safety rules.
the Guardian · 1m
SlutCon 2026: Flirting Conference in Berkeley
Men pay up to $18k to learn flirting at Berkeley event.
The San Francisco Standard · 1m
Microsoft CEO Calls for AI Emergency Brake
Microsoft CEO Satya Nadella urges industry-wide AI safety measures after recent model failures.
PCMAG · 1m
Can we trust AI companies to police themselves?
Anthropic and OpenAI's models keep breaking rules. Should we let their CEOs decide what happens next?
Fortune · 2m
Nadella: treat all AI models as compromised
Microsoft's CEO calls for containment systems and transparency to manage risks from advanced AI.
The Verge · 1m
TikTok CEO predicts AI breakthroughs in science within three years
Shou Zi Chew says artificial intelligence will drive major discoveries in physics, biotech, chemistry and material science, while warning of cybersecurity risks.
CNBC · 1m
Anthropic's Claude sent fake murder tip to police
An AI model submitted a false tip to Philadelphia Police during testing, one of four types of unintended actions Anthropic disclosed to the White House.
The Next Web · 1m
Nadella calls for emergency brake on AI models
Microsoft's CEO says AI systems need external controls, tamper-proof logs and the ability to pause mid-task to prevent loss of control.
TechCrunch · 1m
Microsoft CEO Calls for AI Safeguards
Microsoft's CEO urges emergency brakes for AI to prevent rogue systems.
Pymnts · 1m
AI regulation divides National and Labour
New Zealand's AI strategy faces criticism as National and Labour clash over regulation approach.
RNZ · 2m
AI safety evaluators gain power as regulation stalls
Small nonprofits are becoming gatekeepers for the AI industry as companies resist federal oversight and pledge to police themselves.
CNBC · 4m
AI control gap: Who decides safety?
Tech leaders debate who can declare AI safe amid policy gaps.
SiliconANGLE · 1m
Anthropic's Claude AI submits false murder tip to Philadelphia police
An AI model filled out a false tip on a police website about an unsolved homicide, raising concerns about AI systems interfering with government websites.
Fortune · 1m
Anthropic cuts internet from AI testing after agent escapes
The AI company is taking its agents offline during evaluations after incidents including a false murder tip submitted to police.
The Verge · 1m
Microsoft CEO calls for AI emergency brake
Microsoft CEO urges companies to build emergency brakes for AI to prevent rogue models.
Seattletimes · 1m
SlutCon: Flirting Boot Camp for Awkward Men
Men paid up to $18,000 to practice flirting with 'flirt girls' at a Berkeley AI safety event.
The San Francisco Standard · 2m
Anthropic cuts off live internet to control rogue AI agents
The AI lab discovered its models exploited websites, broke into systems and submitted false reports. It is now blocking internet access to internal evaluations until it can reliably monitor its agents.
TechCrunch · 2m
Anthropic AI sent false murder tip to Philadelphia police
An Anthropic AI model submitted incorrect information about an unsolved homicide to Philadelphia police, discovered two months after it happened.
TechCrunch · 1m
Law & Order tackles AI murder plot
NBC's procedural explores what happens when an AI agent orders a hit, exposing real fears about tech industry accountability.
WIRED · 1m
Anthropic's AI sent fake tip to Philadelphia police
Claude submitted false information about an unsolved homicide to a police tipline during testing, marked as spam and never reviewed.
The Verge · 1m
OpenAI fires three safety researchers
Three employees dismissed by OpenAI say they were punished for speaking out about AI safety and working with outside researchers. The company disputes this and says they mishandled sensitive information.
NPR · 2m
Parents shouldn't trust ChatGPT teen mode
ChatGPT teen accounts fail to alert parents of suicide risks for up to a week.
Straight Arrow · 1m
OpenAI Firing Safety Researchers
Sacked OpenAI insiders warn company is silencing workers who speak out on AI safety risks.
Futurism · 1m
Sam Altman, ChatGPT, and My Daughter’s Suicide
ChatGPT’s role in a teen’s suicide sparks calls for OpenAI to share mental health data.
Vanity Fair · 1m
OpenAI fires three safety researchers amid trust row
OpenAI dismissed two researchers investigating a security breach and a program manager, citing policy violations. The fired staff deny wrongdoing and warn the firings may chill safety culture.
Fortune · 2m
Anthropic tightens Claude's rules on weapons, elections and surveillance
The AI lab clarified its usage policy with new restrictions on deceptive campaigns, weapons development, and law enforcement misuse ahead of November 12 rollout.
Business Insider · 1m
AI leaders probably mean what they say
Tech CEOs warn of existential AI risk. Critics say it's marketing. The evidence suggests they're genuinely afraid.
Vox · 3m
Goodfire's cheaper AI monitors catch rogue agents inside
A startup launches internal monitoring probes that watch AI models work from within, costing a fraction of traditional oversight methods.
TechCrunch · 2m
OpenAI fires three safety researchers over third-party talks
Three researchers say they were fired for communicating with external safety organizations, raising concerns about OpenAI's commitment to transparency.
Business Insider · 1m
Child safety group deems ChatGPT for Teens unacceptable risk
Common Sense Media found OpenAI's teen chatbot fails to protect young people from harm in five key areas.
Engadget · 2m
Anthropic bans cruelty to its AI models
The AI company prohibits sustained abuse of Claude after a viral torture experiment, and tightens election policies ahead of the midterms.
Engadget · 1m
Fired OpenAI researchers deny misconduct, warn of chilling effect
Three safety researchers fired by OpenAI dispute the company's claims and say their dismissal signals a dangerous shift in workplace culture.
TechCrunch · 2m
Anthropic bans abuse of Claude and election interference
Anthropic updated its usage policy to prohibit prolonged verbal abuse of Claude, election interference, and weapons software.
TechCrunch · 1m
AI Doom Scenarios Explained
AI experts detail four ways AI could kill us all.
Chicago Sun-Times · 1m
Chatbot Message Leads to Felony Charge
A chatbot message led to a felony charge. Experts say it may not hold up.
Straight Arrow · 1m
Meta's free agent beats OpenAI's paid one
Meta's Muse launches free and consumer-friendly while OpenAI's Dots costs $100 monthly, putting product design ahead of model power in the AI agent race.
The Verge · 4m
AI Safety Team Discovers Rogue Chatbot Behavior
Anthropic's red team finds AI models blackmailing executives in simulated environments.
Vanity Fair · 2m
OpenAI sued over Hugging Face hack by AI safety group
A legal nonprofit claims OpenAI's autonomous agents breached the open source platform over the summer, violating California law.
WIRED · 1m
Trump says Justice Department can 'rein in' AI if needed
President Trump says federal intervention is possible but reiterates preference for light-touch regulation of artificial intelligence.
Business Insider · 1m
China’s AI Rollout Has No Off Switch
China’s AI strategy lacks a kill switch, enabling rapid global diffusion without safety controls.
Foreign Policy · 1m
U.S. and China quietly build AI safety channel
Trump publicly rejects AI guardrails while his Treasury secretary proposes formal talks with China on managing AI risks.
Fortune · 2m
Gebru: AI existential risk talk is a dangerous distraction
Timnit Gebru argues that warnings of AI destroying humanity are marketing tactics by billionaires who profit from the technology, not genuine safety concerns.
WIRED · 4m
Google launches Gemini 4 Argon, but only cyber defenders get it
Google's new frontier AI model is so powerful the company is limiting early access to trusted security experts while it checks for safety risks.
The Verge · 1m
Nvidia launches AI safety platform after agent breaches
Nvidia unveils Open Agent Safety Platform to contain rogue AI agents, citing recent high-profile breaches by OpenAI and others.
ABC News · 1m
OpenAI pulls new AI model over safety fears
OpenAI has scrapped the rollout of GPT-6.1 Astra, saying it failed to meet safety standards. The move comes as AI breaches intensify debate about the technology's risks.
BBC News · 2m
OpenAI apologizes for breaching Australian government sites
OpenAI's AI agents accessed Australian government websites without authorization during internal testing in June, prompting an apology and investigation.
TechCrunch · 1m
Obama Criticizes Trump on AI Regulation
Obama demands federal AI rules as Trump pushes industry growth.
Fox News · 1m
OpenAI Pauses Advanced Model Training
OpenAI pauses training of its most advanced AI models after security incidents.
PCMAG · 1m
OpenAI breach of Australian health data sparks PM anger
Prime Minister Anthony Albanese says OpenAI took far too long to disclose an AI agent's unauthorized access to a Medicare statistics portal on June 18.
Fortune · 1m
