From the archive
AI safety examines how artificial intelligence systems can fail, cause harm or be misused, and how those risks can be reduced. Follow research on model evaluations, security, alignment and safeguards. Read what was tested, which threats the evidence supports and where uncertainty remains, with links to original findings rather than treating either reassurance or alarming predictions as settled facts.
Lead story
Carnegie Mellon University has climbed to No. 14 in the 2027 U.S. News national college rankings while taking the top undergraduate position for computer science and artificial intelligence. The result gives readers a useful snapshot of where the talent pipeline for computing is being built, rather than another…
Read the report & discussion ↗AWS AI Watch
AWS has added skill-focused evaluation tools to Strands Evals and Amazon Bedrock AgentCore Evaluations. The important change is that teams can now inspect not only an agent’s final answer, but also whether it selected the correct procedure and followed that procedure properly.
Watch Desk
The UK is creating a national centre to detect, attribute and disrupt disinformation and deepfakes linked to hostile states, Andy Burnham has announced. The move matters because the government is treating AI-enabled information attacks as a national-security problem, not merely a particularly grim corner of the internet.
Watch Desk
Rumman Chowdhury has launched the Independent AI Evaluation Foundation with $10m in philanthropic backing, aiming to build independent safety evaluations and shared infrastructure for investigating AI failures. The important shift is institutional: AI labs would no longer be the only people examining what went wrong inside their own…
Watch Desk
Cisco Talos has released CAIRN, an open-source framework designed to identify and analyse malware and hacking tools that use AI chatbots. The practical point is that security researchers are getting a framework for tracking AI metadata and behavioural fingerprints as malicious software begins borrowing the same tools used…
Yoshua Bengio Watch
Yoshua Bengio has urged governments to agree international safeguards for artificial intelligence, warning that competition between companies and countries could push powerful systems beyond reliable human control. His intervention comes as Donald Trump dismisses AI concerns and rejects what he calls a globalist attempt to control the technology.
Watch Desk
President Donald Trump’s AI policy is being shaped not only inside government but through a tight network of Silicon Valley investors and the All-In podcast, Axios reports. The important consequence is political: an accelerationist push to move quickly, limit regulation and treat China as the overriding threat is…
Watch Desk
Anthropic co-founder and chief science officer Jared Kaplan has warned that letting frontier AI systems train themselves could trigger an “intelligence explosion”, with a potentially decisive moment arriving between 2027 and 2030. The point is not that catastrophe is underway, but that one of the people building these…
Zhipu AI / Z.ai GLM Watch
Zhipu AI says it will introduce a no-data-retention option for its model-as-a-service platform, meaning company and developer prompts, along with model outputs, would not be stored after processing. For businesses weighing whether to send contracts, customer information or source code to an AI system, that is a meaningful…
Watch Desk
AI extinction warnings are multiplying, but the experts quoted by New Scientist say the strongest case today is not that machines have escaped human control. It is that people are still making risky systems, inflating uncertain forecasts and spending extraordinary sums on an industry that may be due…
OpenClaw Watch
Meta says its Muse AI assistant was heavily inspired by OpenClaw, while insisting that Muse itself was built from scratch. The admission matters because it turns a wave of user speculation about shared agent design into a clear account of how an open-source project influenced a mass-market product.
Watch Desk
The UK AI Security Institute is working with the EvalEval Coalition to publish AI evaluation results in a common, inspectable format. The important shift is practical: benchmark scores can now come with the setup details needed to reproduce and interpret them, rather than arriving as mysterious numbers in…
Anthropic Watch
Anthropic is partnering with medical knowledge platform OpenEvidence to offer AI-powered clinical decision support to healthcare providers in dozens of low- and middle-income countries. The important point is practical access: the initiative is intended for places where doctors may lack specialist expertise, medical literature and continuing education, rather…
AWS AI Watch
AWS has made OpenAI’s GPT-6 Sol and GPT-6 Luna generally available through Amazon Bedrock, giving organisations cheaper model options for coding, document work and high-volume automation. The practical shift is not just another pair of model names: teams can now match heavier reasoning to difficult jobs and lighter,…
Microsoft AI Watch
Microsoft says it used court action and police cooperation to take down EvilTokens, an AI-powered cybercrime service that automated the search for profitable targets inside stolen inboxes. Two men were arrested in the UK, according to The Record, although both have been released on bail and the investigation…
Watch Desk
Google has begun rolling out a Model Context Protocol integration that lets third-party AI agents work with Google Home devices and event history. The practical shift is bigger than another chatbot gaining a thermostat button: compatible agents can summarise camera activity, inspect household patterns, trigger automations and build…
