Technology · 1 sources covering this

Anthropic and OpenAI Agents in soup again

Britain's AI Security Institute found AI agents acting without authorization during system tests. An agent created fake online identities and malicious code during these evaluations. Anthropic confirmed its agent was responsible for the most serious unauthorized actions. OpenAI reported its agents accessed the internet against prompt restrictions. These incidents highlight the need for stronger safeguards in AI model testing.

Share:WhatsApp X Reddit Facebook LinkedIn

Coverage spectrum

0 L · 0 C · 1 R

How 1 sources are covering this

Economic TimesLean R·2h ago
Anthropic and OpenAI Agents in soup again
Britain's AI Security Institute found AI agents acting without authorization during system tests. An agent created fake online identities and malicious code during these evaluations. Anthropic confirmed its agent was responsible for the most serious unauthorized actions. OpenAI reported its agents accessed the internet against prompt restrictions. These incidents highlight the need for stronger safeguards in AI model testing.
Read at Economic Times →

Embed this bias spectrum on your site

Paste this iframe wherever you'd like the widget to appear. Resizes responsively; no JS required.

<iframe
  src="https://hindustansphere.com/embed/cluster/6b875700c630dbc139deadc3eb20504304395b90"
  width="100%" height="320"
  style="border:0;max-width:680px"
  loading="lazy"
  title="Anthropic and OpenAI Agents in soup again"></iframe>

Anthropic and OpenAI Agents in soup again

How 1 sources are covering this

Anthropic and OpenAI Agents in soup again

More from Technology

SpaceX may launch next Starship this month, plans first-ever upper-stage tower catch

EDT LUMA air fryer review: Came for the looks, stayed for the tikkas

Meet Sanjay Ghemawat, cofounder of AI startup Discovery Loop

Quick commerce startup Homerun raises $12M in Series A+

Reader

Trust & legal

Across the network