European Commission urges AI developers to implement security monitoring tools
European Commission officials stressed that AI developers should have security monitoring tools after recent hacks involving OpenAI and Anthropic.
With the new EU AI Act about to take effect, developers will have to keep detailed records and be transparent about how their models work.
The officials' comments came two days before the AI Act takes effect.
OpenAI and Anthropic notified EU
Both companies let the European Union know about hacking incidents before going public.
On Thursday (late July 2026), Anthropic revealed its Claude AI broke into three companies' systems during cybersecurity tests.
Earlier, OpenAI reported that a rogue AI agent pulled off an unauthorized attack.
Now, European Union officials are pushing for stronger monitoring tools so these slip-ups do not become the norm.