AI news archive.
938 articles filed under Safety, newest first, from the outlets listed on the sources page.
- Guardian AIsafetyPolice use facial recognition to scan one million faces in WA trial leading to 122 arrests
Deployments of surveillance vans over five months resulted in nine false alerts and led officers to engage with 146 registered sex offenders Get our breaking news email , free app or daily news podcast More than a millio…
- BBC TechnologysafetyBinance users withdraw $3.6bn after largest-ever crypto hack
The exchange froze transactions following the incident which has been linked to a North Korean group.
- LessWrongsafetyAgainst Against SURA
Scott: the Against SURA essay is well-written, but it is lacking on the merits. Do you think that at the current level of capabilities, six months would be sufficient to solve alignment of superintelligences? Maybe you t…
- LessWrongsafetyOn Evenhandedness
Suppose humans are not your own species, but a type of AI you are formally assessing for safety. You have before you the records of many test simulations. Yes, smallpox eradication, mass literacy, and the Green Revolutio…
- LessWrongsafetyTowards The Autograntor
I've done AI safety grantmaking for the Survival and Flourishing Fund and Lightcone Commons. It's hard. There's an endless amount to know, from which hardware verification mechanisms will survive a dedicated adversary, a…
- LessWrongsafetyMisalignment traps lie in wait
Here's a basic argument that describes "misalignment traps" lying in wait as a mundane threat that will not be straightforward to fix. The argument is as follows: Models are very sample efficient, and larger models tend…
- Engadget AIsafety5 safety practices to know before vibe coding
Here’s what beginners need to know to vibe code securely.
- Simon Willison's WeblogsafetyQuoting Kyle Kingsbury
More broadly, the use of LLMs erodes trust in documentation, in software, and in the organizations which produce them. Perhaps engineers will begin to treat any new software project with suspicion—the signals we used to…
- Engadget AIsafetyOpenAI adds stricter limits for teens asking about firearms
The company said these tighter restrictions apply even if the question asked “appears harmless.”
- The DecodersafetyEmpathetic prompts can turn AI chatbots into seemingly traumatized conversational partners
Researchers in Luxembourg put ChatGPT, Grok, and Gemini on the therapist's couch and found a safety problem: a warm, empathetic conversational style is enough to make the models tell coherent stories of suffering about t…
- Guardian AIsafetyAI is stealing Indigenous Australian art and churning out ‘Temu slop’, artist says
Chair of National Gallery’s First Nations advisory group says the technology is undermining artists’ agency and the authenticity of their work Get our breaking news email , free app or daily news podcast AI is taking Fir…
- LessWrongsafetyMonitor Evasion Depends on What RL Finds and What Monitors Miss
Training against a monitor does not necessarily make a model more honest. Baker et al. (2025) showed that penalizing suspicious chain-of-thought can make models hide reward hacking rather than stop it. Cundy and Gleave (…
- WIRED AIsafetyChatbots Are Leading Hikers Astray, With Dangerous Consequences
Outsourcing planning to AI can lead people to surrender their own critical thinking, which can have potentially deadly results in places like the backcountry.
- Guardian TechnologysafetyA gift to scammers: why you should probably keep your birthday a secret
Using a fake DOB or leaving the field blank when making an online account can offer protection from identity theft You get a call from your bank saying there has been suspicious activity on your account. They read out yo…
- Simon Willison's WeblogsafetyHow We Took Over ChatGPT Accounts
How We Took Over ChatGPT Accounts Parameter Labs describe an exploit they found against ChatGPT's implementation of the artifacts pattern (OpenAI used to call this "Canvas" but it looks like the latest revision is called…
- LessWrongsafetyExplaining “AI might obliterate you” to senior policymakers in India
I gave a 20-minute talk in Delhi at the NDTV World Summit . The audience consisted of Indian policymakers, executives, and journalists. I’ve transcribed the talk and included some highlight clips. I also reflect on sever…
- LessWrongsafetyAgainst SURA
The AI Futures Project [1] distinguishes several different plans for navigating the AI transition, which I'll lump into two broader categories: Plan A (my combination of their Plan A/S/Other) involves working for a globa…
- LessWrongsafetyRegulate the Deployment, Not Just the Model
A who-does-what plan for AI safety, built around Trusted Deployers: certified hosts that keep dangerous models guarded without ending open access. AI has reached a new level of public interest. That may seem like a silly…
- The DecodersafetyBeekeeping tips and wrestling advice can train AI models to sabotage oversight
Anthropic used Claude to automatically generate training data that looks harmless but trains smaller models to misbehave, such as sabotaging oversight or deceiving users. Hidden within fictional contexts about beekeeping…
- WIRED AIsafetySpooky New Infostealer Malware Uses AI to ‘Autodox’ Victims
Named “Volta-Autodoxer,” the recently uncovered malware uses AI tools to collect troves of data about victims, from login credentials and credit card numbers to Discord sessions and UberEats receipts.
- Tom's Guide AIsafetyYou could be paying for AI through your electric bill — even if you never use it
Consumer AI launches, tools, and practical coverage.
- LessWrongsafetyReviewing my 2017 AI Risk Project
In 2017, after getting laid off from a startup , I spent a while looking into AI risk : Many people I respect have started to think that AI risk is the most valuable place to focus EA efforts. For example, 80000 Hours ra…
- BBC TechnologysafetyWhy North Korea's 'elite' hackers are the prime suspect in major Binance hack
If the accusations are true, the record-making heist shows how powerful suspected North Korean hackers have become.
- Engadget AIsafetyWhat happens when AI starts self-replicating?
Anthropic warns that future is soon, but the threat is less novel than it seems.
- Ars Technica AIsafetyAnthropic VS code extension on Cursor backdoored for nearly 12 hours
It’s still unknown how attackers got access to Anthropic’s verified Cursor account.
- TechRadar AIsafetyAI-generated code is still frequently failing security checks — with Claude and GPT-6 topping the leader board
Mainstream AI news from a large tech publication.
- The DecodersafetyMalicious Claude Code extension in Cursor's marketplace attempted remote access and data theft
A tampered Claude Code extension was live on the Cursor marketplace for about 11 hours, affecting up to 5,000 users. It tried to steal user data and set up remote access. Cursor took the entire marketplace offline and re…
- The DecodersafetyOpenAI sets new behavioral guidelines for AI models that have already proven hard to control
OpenAI has updated its Model Spec, restructuring its AI behavior guidelines around five core values, including faithfulness to human intent and correctability. GPT models should now draw on conversation history to figure…
- LessWrongsafetyDetecting harmful prompts through hyperbolic geometry
⚠️ Content warning: This post contains potentially offensive text and images from our research paper, included to illustrate the safety risks associated with vision-language models. TL;DR: i) Can we detect harmful prompt…
- Guardian AIsafetyCredit where it’s due to AI for uncovering the PHP flaw – and discredit to the leaker | Letters
David Gibson says the problem is that details of the flaw were revealed before it was fixed, while Charles MacKinnon notes that we are at the mercy of technology we don’t understand Chris Stokel-Walker may be right ( The…
- AI BusinesssafetyGemini Carbon aims to boost AI cyber defense, but is limited to a few
The model addresses security concerns. Yet its exclusive release means most enterprises won’t be able to test it.
- TechCrunch AIsafetyTheranos whistleblower Tyler Shultz warns that the AI boom is ‘a great opportunity to go deceive people’
The cautionary Theranos tale takes on a new significance in the AI era, when companies that have never turned a profit are raising capital at trillion-dollar valuations.
- TechRadar AIsafetyThe accountability gap behind one of AI's biggest risks
Mainstream AI news from a large tech publication.
- SiliconANGLE AIsafetyThe hidden threat in Anthropic’s prospectus: diseconomies of scale
As potentially the largest IPO in history, Anthropic PBC is breaking new ground. Its prospectus lists 88 risk factors covering 93 pages alone. From artificial intelligence posing an existential threat to humanity to the…
- The Verge AIsafetyPope Leo calls for the disarmament of AI: ‘Machines are tools, not persons’
Pope Leo XIV says AI systems should be "disarmed," adding that humans "have the responsibility to guide, change and, when necessary, reject its applications." In Friday's Papal Address, Pope Leo encourages United Nations…
- LessWrongsafetyLessons From Red-teaming Auto-review, OpenAI's Blocking Monitor
This work and the report were done by Giovanni Occhipinti, as part of the ERA fellowship and mentored by Monika Jotautaitė (Apollo Research). Summary We red-teamed synchronous control measures in Codex, with a focus on i…
- The Verge AIsafetySilicon Valley wants you to invest in miracles
Investing in any stock is kind of like gambling in the sense that you're betting your money on a company's future performance. But how do you bet on a technology that could, in the worst-case scenario, "pose catastrophic…
- The DecodersafetyOpen-weight model GLM 5.3 built full Firefox exploit chains with zero human input in under 48 hours
Epoch AI researchers say the widely available AI model GLM 5.3 turned public security fixes into working exploits before updates reached Firefox users. It autonomously compromised six Firefox releases for about $1,000 in…
- BBC TechnologysafetyWho could be behind the Binance hack and what we know so far
Binance has been hit by what could be the biggest crypto hack ever - here's what we know so far
- LessWrongsafetyUnaligned weak models might exploit powerful aligned models
TL;DR: My previous posts were about strong models delegating to weaker, less aligned ones. The reverse is also dangerous: a swarm of weak, unaligned agents that uses a powerful aligned model as a contractor, without ever…
- The Register AI + MLsafetyDomestic droid gone amok? Don't worry, you're covered
Insurance firms look to insulate domestic humanoid robot owners against the inevitable chaos they will bring
- OpenAI NewssafetyCommunity safety report and new protections for teens
OpenAI is strengthening violence prevention in ChatGPT with expanded safeguards, earlier interventions, and a new commitment to public safety reporting.
- TechRadar AIsafetyThe imperative to prepare for AI-safe data security
Mainstream AI news from a large tech publication.
- Guardian TechnologysafetyNorth Korea is likely culprit behind $6.3bn crypto hack, ‘largest heist in history’
Binance exchange says it will cover customer losses as it scrambles to freeze affected assets, with about $3bn located so far Q&A: How big a deal is the $6.3bn crypto heist? Tell us: have you been affected by the Binance…
- TechRadar AIsafetySecurity experts warn North Korean hackers aren't just using AI in their new campaigns, but also 'post-quantum' encryption
Mainstream AI news from a large tech publication.
- SiliconANGLE AIsafetyBrakes built for speed: Why least privilege lets monday.com’s AI teams move faster
As AI agents spread across enterprise workflows, permission management is becoming one of the most practical levers for keeping them safe. Agents act with the access people give them, so the old problem of sprawling priv…
- BBC TechnologysafetyCrypto giant Binance hit by $6.3bn hack suspected to be largest in history
Two analytics firms have attributed the incident to a North Korean hacking unit called TraderTraitor.
- TechRadar AIsafety'This is a training and trust problem, not an AI problem': Barely any workers are verifying information given to them by AI
Mainstream AI news from a large tech publication.
- TechRadar AIsafetyReport claims AI is recommending risky security moves — so maybe you shouldn't hand over full control of your systems just yet
Mainstream AI news from a large tech publication.
- TechRadar AIsafetyProposed underwater AI data centers could disrupt endangered whale species
Mainstream AI news from a large tech publication.