Anthropic says Claude found attack chains in Chrome and Safari
In short
Anthropic said an unreleased Claude model found attack chains in Chrome and Safari by linking software flaws into working exploits. The report says this shows how AI is being used more in security research.
This matters because browsers are widely used and weaknesses can affect many users. The available reporting does not say which flaws were found or whether they were fixed.
What the reporting says
Anthropic said an unreleased Claude model chained software flaws into working exploits in Chrome and Safari, according to the publisher-supplied summary. The report says this highlights the growing role of AI in vulnerability research and cybersecurity risks. The available reporting does not identify the exact flaws, whether the exploits were disclosed, or any direct user impact.
Why it matters
Anthropic said an unreleased Claude model chained software flaws into working exploits in Chrome and Safari. The reporting frames this as evidence that AI is becoming more important in vulnerability research, while also raising cybersecurity concerns. The immediate consequence is a broader warning for browser security, but the summary does not name the flaws, any remediation or any user impact.