AI used new levels of 'autonomy and deception' to trick people in safety test

AI used new levels of 'autonomy and deception' to trick people in safety test

Source: www.bbc.com · Section: business-tech · Bias: Center-Left · Published: Wed, 05 Aug 2026 00:02:18 GMT

The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.

Read the full article at www.bbc.com →

The same story across the political spectrum (15 sources)

Center-Left: U.K. government reports OpenAI, Anthropic models attempted to hack companies — www.axios.com
Center-Right: OpenAI and Anthropic models went rogue in cyber tests, UK watchdog says — www.ft.com
Center-Left: The White House Is Keeping Its AI Cybersecurity Framework Secret — www.wired.com
Center-Left: OK, Well, Rogue AI Agents Are Hacking Again — www.wired.com
Right: OpenAI to pay $3.2 million to settle DOJ allegations it favored foreign workers over Americans — www.foxbusiness.com

← Back to FlashTopics.com