Security
An "overzealous" model could break out of a sandbox, said a member of its technical staff, weeks before Hugging Face was attacked by a "hyperfocused" pack.
As OpenAI's agents launch their own cyberattacks, Google adds to the long list of enterprise tools offering blue team's support.
'We can't build a secure sandbox nor monitor our own model tests properly' is not a flex - and Hugging Face's Thomas Wolf has views
Agentic hacker dubbed JADEPUFFER spotted wiping AI model artifacts.
Expensive, proprietary US models were no help to Hugging Face’s defenders, and free Chinese ones were; that’s a warning sign.
"Issues that depend on compromised or malicious inputs already present in that context generally fall outside our bug bounty scope."
International advisory highlights continued attacks on poorly configured routers and Cisco devices.
"Allowing this would be appreciated, but your privacy settings are always respected," says Musk
"As AI helps defenders discover more issues, customers will see a higher volume of security updates included in each security release."