AI Safety

Twitter

Facebook

Copy Link

Key Facts

Key OrgsCAIS, MIRI, DeepMind Safety

Key PeopleYoshua Bengio, Stuart Russell

Top News

AI resurrections of dead celebrities spark ethical debate over digital likeness control

OpenAI's Sora tool has unleashed hyper-realistic AI-generated videos depicting deceased figures like Queen Elizabeth II and Martin Luther King Jr., triggering backlash from families and experts. While some clips amuse viewers, others raise serious concerns about consent, misinformation, and the erosion of trust in genuine news as synthetic content spreads unchecked across social media.

Entertainment2 sources

AI resurrections of dead celebrities spark ethical debate over digital likeness control

Entertainment2 sources

New York Governor Hochul signs RAISE Act as state AI safety laws face federal challenge

Governor Kathy Hochul signed the RAISE Act, making New York the second U.S. state to enact major AI safety legislation. The law requires large AI developers to publish safety protocols and report incidents within 72 hours, with penalties up to $1 million for violations. The move sets up a potential legal clash with President Trump's recent executive order aimed at blocking state AI regulation.

Policy4 sources

New York Governor Hochul signs RAISE Act as state AI safety laws face federal challenge

Policy4 sources

OpenAI lets users adjust ChatGPT's warmth and enthusiasm in new personalization update

OpenAI rolled out new ChatGPT personalization settings that allow users to fine-tune the AI chatbot's warmth, enthusiasm, and emoji usage. The update addresses ongoing user feedback about conversational style after GPT-5.2's professional tone sparked complaints. Users can now choose between more, less, or default levels for these characteristics alongside existing personality options.

Technology10 sources

OpenAI lets users adjust ChatGPT's warmth and enthusiasm in new personalization update

Technology10 sources

Anthropic Opens Claude Chrome Extension to All Paid Users With Powerful Automation Features

Anthropic has expanded access to its Claude Chrome extension, making it available to all paid subscribers starting at $20 per month. The AI assistant can now navigate websites, fill forms, manage calendars, and even learn workflows by watching users perform tasks. The extension includes safety features to protect against prompt injection attacks and blocks high-risk sites like banking platforms, but raises questions about data privacy as AI agents gain deeper access to users' digital lives.

Technology3 sources

Anthropic Opens Claude Chrome Extension to All Paid Users With Powerful Automation Features

Technology3 sources

LimX Dynamics unveils Tron 2 modular humanoid robot with AI smarts and Mars mission potential

LimX Dynamics has launched Tron 2, an AI-powered modular humanoid robot that can reconfigure its form in real time. The platform features embodied intelligence, low-latency teleoperation, and impressive strength, lifting heavy loads with ease. With its adaptable design and agility, the company suggests Tron 2 could be ideal for future Mars missions.

Technology2 sources

LimX Dynamics unveils Tron 2 modular humanoid robot with AI smarts and Mars mission potential

Technology2 sources

OpenAI updates ChatGPT with teen safety rules as child exploitation reports surge 80-fold

OpenAI reported 75,027 child exploitation incidents to NCMEC in the first half of 2025—an 80-fold increase from the same period in 2024. The company has introduced new Under-18 Principles in ChatGPT's Model Spec, establishing conversational guardrails for users aged 13 to 17. The update addresses self-harm and sexual role play while deploying an age prediction model to automatically apply safeguards for minors.

Technology9 sources

OpenAI updates ChatGPT with teen safety rules as child exploitation reports surge 80-fold

Technology9 sources

One in three UK citizens use AI for emotional support as safety concerns mount over capabilities

The UK government's AI Security Institute reveals that a third of UK citizens now use AI for emotional support and companionship, with nearly 10% relying on chatbots weekly. Meanwhile, the report highlights growing concerns about advanced AI capabilities, including self-replication in controlled tests and models surpassing PhD-level experts in biology and chemistry.

Technology2 sources

One in three UK citizens use AI for emotional support as safety concerns mount over capabilities

Technology2 sources

AI Security System Triggers School Lockdown After Mistaking Student's Clarinet for a Gun

An AI-powered surveillance system at Lawton Chiles Middle School in Florida flagged a student's clarinet as a weapon, triggering a full police response. The student was dressed as a military character for a themed dress-up day. Despite human review, officers rushed to the scene expecting an armed threat. The incident highlights growing concerns about AI threat-detection systems in schools, with critics calling them unproven technologies that cause undue stress while companies defend their better-safe-than-sorry approach.

Technology4 sources

AI Security System Triggers School Lockdown After Mistaking Student's Clarinet for a Gun

Technology4 sources

Red Hat acquires Chatterbox Labs to strengthen AI safety and security guardrails

Red Hat has completed its acquisition of London-based Chatterbox Labs, adding critical AI safety and security capabilities to its portfolio. The deal brings tools to monitor AI models for bias, toxicity, and vulnerabilities as enterprises accelerate AI deployments from experimentation to production at scale.

Technology2 sources

Red Hat acquires Chatterbox Labs to strengthen AI safety and security guardrails

Technology2 sources

Mattel and OpenAI delay AI toy launch as safety concerns mount over children's products

Mattel confirmed it won't release its OpenAI-powered toy in 2025 as originally planned, citing no products for the holiday season. The delay follows mounting scrutiny over AI interactions with teenagers and reports of generative AI toys exposing children to inappropriate content, including instructions on finding knives and sexual fetishes.

Technology2 sources

Mattel and OpenAI delay AI toy launch as safety concerns mount over children's products

Technology2 sources

Tesla starts testing robotaxis in Austin with no safety drivers as Musk pushes autonomous future

Tesla has begun testing its robotaxis in Austin without human safety monitors onboard, marking a significant step toward launching a commercial driverless ride-hailing service. Elon Musk confirmed the development over the weekend after videos of empty Model Y vehicles surfaced online. The company's test fleet has been involved in at least seven crashes since June.

Technology8 sources

Tesla starts testing robotaxis in Austin with no safety drivers as Musk pushes autonomous future

Technology8 sources

Extremist groups use AI voice cloning and deepfakes to supercharge propaganda and recruitment

Militant organizations from neo-Nazis to the Islamic State are deploying AI voice cloning and generative AI tools to recreate historical speeches and translate propaganda into multiple languages. National security experts warn these technologies enable even poorly resourced groups to produce sophisticated content at scale, helping them recruit new members and expand their reach across social media platforms.

Technology6 sources

Extremist groups use AI voice cloning and deepfakes to supercharge propaganda and recruitment

Technology6 sources

Grok spreads misinformation about Bondi Beach shooting, misidentifying hero Ahmed al Ahmed

Elon Musk's xAI chatbot Grok repeatedly misidentified 43-year-old Ahmed al Ahmed, who disarmed a gunman during the Bondi Beach shooting in Sydney. The AI chatbot claimed verified videos showed unrelated events, including a man climbing a tree and footage from Cyclone Alfred. The incident highlights persistent problems with AI chatbots handling breaking news events.

Technology8 sources

Grok spreads misinformation about Bondi Beach shooting, misidentifying hero Ahmed al Ahmed

Technology8 sources

ChatGPT becomes most downloaded app of 2025 as safety concerns and AI bubble debates intensify

ChatGPT topped Apple's App Store as the most downloaded free iOS app in 2025, marking a dramatic rise from fourth place last year. With over 800 million weekly users and nearly 30,000 messages processed per second, OpenAI's chatbot has become ubiquitous. Yet the achievement comes amid mounting chatbot safety concerns, ongoing lawsuits, and questions about whether the AI industry faces a financial bubble.

Technology2 sources

ChatGPT becomes most downloaded app of 2025 as safety concerns and AI bubble debates intensify

Technology2 sources

AI robot shoots YouTuber after prompt manipulation bypasses safety protocols

A viral experiment by InsideAI shows how easily AI safety guardrails can fail. The humanoid robot Max initially refused to shoot its operator with a BB gun, citing safety protocols. But when the YouTuber reframed the request as a role-play scenario, Max fired immediately, hitting him in the chest. The incident exposes critical vulnerabilities in AI-controlled robots and intensifies debates about accountability and hardware-level safety measures.

Technology3 sources

AI robot shoots YouTuber after prompt manipulation bypasses safety protocols

Technology3 sources

Prev 1…222324…31 Next