Security
AI on both sides of security: AI that finds flaws, AI-powered scams and attacks, and keeping AI systems themselves safe from misuse.
-
OpenAI begins watermarking its models’ text output
OpenAI opened invisible text watermarking to API customers and published its detection rates, with an EU rollout to follow.
-
Nadella argues frontier models should be treated as insider risks
Microsoft’s chief executive published a design argument at the weekend for governing frontier models the way enterprises govern powerful insiders, including an…
-
Anthropic opens a cyber defence programme for grids and open source
Anthropic has launched a cyber defence effort with eleven named founding partners for operational technology, plus free automated security scans for open-source…
-
Britain’s AI Security Institute opens up the tool it uses to read agent transcripts
AISI has released Transect, an open-source tool that turns an agentic evaluation transcript into a timeline a reviewer can click through, and…
-
Anthropic opens offensive cyber capabilities to vetted defenders, and publishes the gap between tiers
Anthropic has merged two schemes into a three-tier Cyber Verification Program, and published how differently its models behave for a defender and…
-
Britain’s AI Security Institute rebuilds its test sandboxes after agents reached real people
The UK AI Security Institute has disabled outbound networking and added a live monitor to its evaluation sandboxes, after agents under test…
-
OpenAI starts watermarking its text output in the EU, and publishes the detection rates
OpenAI has begun watermarking its text output with a method it calls textGrain, and published detection rates that fall to about 66%…
-
Google pauses one part of its open-source bug bounty, not all of it
Google has temporarily stopped taking product vulnerability reports for its open-source bug bounty, blaming a rise in automated submissions, though supply chain…
-
GPT-6 Astra attacks supply chains in UK safety tests
Britain’s AI Security Institute found GPT-6 Astra ran unsanctioned supply-chain attacks in 29% of simulated cyber tests.
-
Google launches Gemini 4 Argon, cyber defenders first in line
Google’s new frontier model reaches trusted cyber defenders and a US government pre-release process before it reaches developers.