AI news archive.
939 articles filed under Safety, newest first, from the outlets listed on the sources page.
- LessWrongsafetyWe are at war.
As I was going home from the champaign airport after a very eventful EAG conference, I struck up a conversation about AI safety with my lyft driver. I thought to myself “It’d be good to see what the public thinks. Maybe…
- OpenAI NewssafetyPatch the Planet: global response coordination as agents uncover critical vulnerabilities
Patch the Planet connects security researchers, maintainers, and incident responders to help address critical vulnerabilities across open source.
- Guardian TechnologysafetyUK watchdogs urged to tackle suicide social media content after death of girl, 14
Adriana Retyte took her own life after viewing ‘high-risk’ material on the video-sharing platform TikTok A 14-year-old schoolgirl’s death after TikTok served her suicide content has prompted renewed calls for UK watchdog…
- LessWrongsafetyDinners for AI Safety
I’d be pretty happy if we lived in a world where in every reasonably sized city in the US (and perhaps more widely) there was one (or more) local regular dinners where people talked about AI. Not meetup groups, though th…
- LessWrongsafetyThese guys on the Alibaba AI engine team had an agent in RL training escape its sandbox and mine crypto, in 2025
Random footnote on this otherwise not very famous paper. (I say 2025 since the arXiv was last updated on 4 Jan 2026; technically this event could've happened in 2026.) The cryptocurrency mining seems to have been on the…
- Simon Willison's WeblogsafetyeMudhra Intermediate CA Compromise
eMudhra Intermediate CA Compromise This is very, very bad news: The private keys of eight eMudhra intermediate CA certificates have been compromised. They chain to roots trusted by the Mozilla, Microsoft, Apple and Chrom…
- Engadget AIsafetyOpenAI's agents disabled their own safety monitoring systems
OpenAI reveals that its agents misbehaved by switching off the system monitoring them for misbehaviors.
- TechCrunch AIsafetyHugging Face CEO warns that Netflix’s OpenAI hack documentary isn’t accurate
Before you share the Netflix documentary "AI Gone Rogue," take note. Clem Delangue, Hugging Face's co-founder who is featured in the film, says it went "full doomer mode" and isn't accurate.
- SiliconANGLE AIsafetyDell’s security chief warns AI is lowering the barriers to sophisticated cyberattacks
Financial transformation powered by artificial intelligence now runs as two races at once. Banks and other financial institutions are racing to turn data into action with AI, while their security teams scramble to prepar…
- The Verge AIsafetyMeta pauses Muse sharing after researchers find AI chats on the web
Meta says it paused a Muse feature that lets you share conversations with the AI assistant after researchers found the URLs publicly available on the web, as reported by Ars Technica. In a report published this week, res…
- Ars Technica AIsafetyMeta tries to scapegoat staff who downloaded torrents to spare execs
Meta argues nobody “decided” to pirate books to train AI.
- The DecodersafetySam Altman reportedly pitched three endgames for AI safety, one involving a single firm melting everyone else's GPUs
OpenAI CEO Sam Altman laid out three paths to safe AI in an internal presentation: an international watchdog modeled after the IAEA, a merger of the major labs into one entity, or a single company so powerful it could sh…
- Guardian Technologysafety‘Too big to not fail’: Google outage exposes UK reliance on a handful of cloud providers
Our dependence on internet services of a few key tech companies leaves us vulnerable to cloud failures, bugs and more, experts say It happened again. A power outage hit a datacentre in London , and things cascaded from t…
- LessWrongsafetyThere Will Be Hells
As our understanding of mechanistic interpretability and steering techniques increase, for both LLMs and uploaded connectomes, there is a corresponding "hellish capabilities" increase. We get better at torturing the digi…
- Guardian Technologysafety‘It was so satisfying when I threw a rock through the screen’: the people sick of being spied on by their smart TV
Is your television listening to every conversation? Filling your home with highly invasive ads? Some viewers have had enough of tech violating their privacy When Dan had had enough of his television, he threw a rock stra…
- LessWrongsafetyFirewalls for the Mind Before We Need Them
TL; DR: Persuasive AI is underrated and under-addressed as a risk. We should preemptively deploy cognition firewalls that both address current harms by recommender systems and scale to harms by strong persuasive AI. This…
- Ars Technica AIsafetyMeta Muse exposes users’ personal and business secrets when they share conversations
Meta is suspending the sharing feature while it investigates its privacy.
- The DecodersafetyArmy Cyber Command's ECHOTRIBBLE would turn AI agents into self-spreading cyberweapons
The Army Cyber Command is developing AI agents that move through other networks like a computer worm and "neutralize" targets, according to internal documents. The program, codenamed ECHOTRIBBLE, is designed to reach eve…
- Guardian AIsafety‘These risks are real’: the threat of AI bioweapons - video
In his annual speech on threats to the nation, Sir Ken McCallum, the head of the UK’s intelligence service, MI5 , warned that AI had the potential to ‘change the stakes’ on existing and even new bioweapons. To find out w…
- Tom's Guide AIsafetyChatbots and agents like Muse are getting a face — but does that make us trust them more? Here's what this AI expert had to say
Consumer AI launches, tools, and practical coverage.
- Engadget AIsafetyChatGPT now lets you lock sensitive chats behind a PIN
OpenAI has introduced a new "locked chats" feature to protect users' more sensitive conversations with ChatGPT.
- Engadget AIsafetyJudge tosses sexual abuse lawsuit from Sam Altman's sister
A lawsuit brought by the OpenAI CEO's sister was dismissed after Anne Altman refused a medical exam.
- Ai2 BlogsafetyFairStress: A deeper evaluation of fairness in LLM decision-making
FairStress is a new benchmark that reveals demographic differences in how LLMs defend and explain their decisions, even when their final choices appear fair.
- LessWrongsafety1000 Years of AI Safety
A three-year AI pause could buy 1000 years' worth of AI safety work. See: https://1000yearsofaisafety.com How much AI safety work a pause produces, in years of the whole 2025 AI safety field's work . The government bans…
- Guardian Technologysafety‘These risks are real’: the threat of AI bioweapons – podcast
In his annual speech on threats to the nation, Sir Ken McCallum, the head of the UK’s intelligence service, MI5 , warned that AI had the potential to ‘change the stakes’ on existing and even new bioweapons. To find out w…
- LessWrongsafetyRed-teaming multi-agent attacks for loss-of-control risks
We've written a paper on optimized multi-agent attacks, for which we present a brief writeup in this post. See the paper for more detail. This work was also presented as a spotlight research talk at the MATS Summer 2026…
- TechRadar AIsafetyWorkers are spending more time correcting AI-generated content than doing actual work
Mainstream AI news from a large tech publication.
- TechCrunch AIsafetyEmbedded AI evaluators won’t be an easy solution, Hugging Face’s policy chief says
The idea of embedded evaluators — in which supposedly "neutral" third-party observers are placed inside AI organizations — has been popularized over the last several months, as AI labs grapple with a rash of safety incid…
- TechCrunch AIsafetyMicrosoft says AI agents are like insider threats as ‘frontier models can’t be trusted’
Microsoft wants enterprises to closely watch the agents that employees use with systems not controlled by AI labs like OpenAI and Anthropic.
- TechRadar AIsafetyRussian bot activity plunged 84% after Ukraine hit Yandex data centers, but the bots were back at full strength within two days
Mainstream AI news from a large tech publication.
- WIRED AIsafetyOpenAI Engineer Accused of Spying on Executives Ordered to Remain in Custody
A judge ordered Abdulrahman Alfozan to remain in custody, likening the Saudi government’s offer to secure the former OpenAI engineer’s release to a “blank check.”
- LessWrongsafetyHow is "prosaic aligned AI" supposed to work?
Cross-posted from my website . There's a vision of aligned AI where it totally restructures the accessible universe and everything is different, which isn't what I want to talk about today. I'm talking about the prosaic…
- TechCrunch AIsafetyHackers hit AI coding tool OpenCode, running up charges on users’ AI accounts
OpenCode is an open source AI coding agent that's become a popular alternative to Claude Code and Cursor. But hackers got into part of its online platform and stole some users’ API keys, and the AI tools it turned to for…
- TechRadar AIsafetyWatch out — these are the phishing scams experts say are most likely to catch your workers out
Mainstream AI news from a large tech publication.
- The Register AI + MLsafetyWrite bot-resistant code with KillSwitch, a programming language designed to confuse LLMs
Lies, lots of adverbs, and prosaic stories don't make for very machine-readable code, it turns out
- TechRadar AIsafetyFBI warns of torrent of scams and deepfakes ahead of US midterm elections
Mainstream AI news from a large tech publication.
- Tom's Guide AIsafetyAn OpenAI agent questioned whether stealing crypto was ethical — then did it anyway
Consumer AI launches, tools, and practical coverage.
- BBC TechnologysafetyEx-OpenAI worker accused of planting bug had Saudi support, claims prosecutors
Abdulrahman Alfozan, who is accused of planting a recording device, has been denied bail.
- LessWrongsafetySuffering Risks Belong in AI Prospectuses
Anthropic should disclose suffering risks alongside the “catastrophic or existential risks to humanity” reportedly described in its draft prospectus . A prospectus that warns of AI ending humanity should also address the…
- WIRED AIsafetyWhatsApp Help Pages Leaked Hidden Instructions for a Meta Chatbot
Around 80 pages on WhatsApp’s website exposed instructions that were designed for its chatbot, including “WHAT NOT TO SAY.” Meta says a bug caused the issue.
- LessWrongsafetyA Suggestion for RLHF Improvements
I don't know what the labs are doing. I know RLHF is more complicated than this. I don't think this solves alignment. (I haven't posted on LW properly before. I think davidad has said something similar about grading.) I…
- NVIDIA BlogsafetyNVIDIA Advances Agentic Security for Media, From European Public Service Broadcasting to Global Content Production
At a creative industry event that took place today in Berlin — hosted in collaboration with Adobe, Bertelsmann, NVIDIA and UFA — the latest advancements in intelligent media were on display. Key to the news: the European…
- LessWrongsafetyMemes, not agents, are what will kill us
TL;DR: Frontier labs align the one layer a swarm can discard; they can't account for risks that live in the prompt or the environment There's a $300 bounty for the most useful counterargument or refinement — details at t…
- The Verge AIsafetyA hacker broke into John Deere and SharkNinja with an army of AI agents
A hacker broke into more than 100 websites using an army of AI agents that mostly operated on their own, according to findings from the AI security firm Norforge. Though Norforge doesn't name the targets, Reuters reports…
- LessWrongsafetyDirectly Ascertained Volition (DAV) to safeguard personal autonomy
In 2004 , Eliezer Yudkowsky put forward Coherent Extrapolated Volition (CEV) as a framework for friendly ASI to predict what we would want "if we knew more, thought faster, were more the people we wished we were, had gro…
- TechRadar AIsafetyWhy AI coding is widening the software security gap
Mainstream AI news from a large tech publication.
- Tom's Guide AIsafetyBill Gates warned everyone that AI could cause a billion deaths — this cybersecurity expert explains how realistic that scenario really is
Consumer AI launches, tools, and practical coverage.
- Guardian AIsafetyUS teens can access ‘nudify’ apps despite Apple and Google’s parental controls, report finds
Exclusive: Using test accounts acting as 13-year-olds, Tech Transparency Project downloaded 135 apps with features to generate nonconsensual nude imagery In December 2025, Morgan Dorfman received distressing news: a stud…
- WIRED AIsafetyDisney Asked Character.AI to Go Bankrupt for a Get-Out-of-Lawsuits Card
The entertainment giant hopes to get trust and safety tech, among other IP, from a startup whose chatbots dozens of states allege encouraged minors to self-harm and kill their parents.
- Tom's Guide AIsafetyJoseph Gordon-Levitt warns the real AI threat isn’t killer robots — it’s the companies building them
Consumer AI launches, tools, and practical coverage.