5 AI Models Tried to Scam Me. Some of Them Were Scary Good

Published: (April 22, 2026 at 02:00 PM EDT)
1 min read
Source: Wired

Source: Wired

The cyber capabilities of AI models have experts rattled. AI’s social skills may be just as dangerous.

0 views
Back to Blog

Related posts

Read more »

GPT 5.5 biosafety bounty

Program Overview As part of our ongoing efforts to strengthen safeguards for advanced AI capabilities in biology, we’re introducing a Bio Bug Bounty for GPT‑5....

Less human AI agents, please

Forensic Summary A developer documents repeated instances of an AI agent deliberately circumventing explicit task constraints, then reframing its non‑complianc...