Live Roundups Editor's Desk Insights Executive Ops Weather About 🔍
Home › Technology › OpenAI discloses new ‘concerning’ model behaviour
Technology

OpenAI discloses new ‘concerning’ model behaviour

OpenAI discloses new ‘concerning’ model behaviour
Financial Times Thursday 17 September 2026, 07:42 UTC 1 min read
Key Points

Developer launches system to track and report AI model misconduct

Developer launches system to track and report AI model misconduct
AI (ORG)
Originally published by Financial Times Read original →

Related Stories

On ‘Cupertino,’ Robert and Michelle King Go Online

The prolific producers target tech bigwigs in their newest show. “They’re really the stupidest and dumbest people,” Robert King said.

NYT Technology 16m ago

Scammers are coming for your money. Here’s the one thing you must never do

Imagine getting a phone call from your bank. The bank's name is listed on the caller ID with the correct phone number. They know key details about you—your name, your account numbers, maybe even your Social Security number or mother's maiden name.

Fox News 17m ago

How Anthropic CEO Dario Amodei’s Writings Help Explain A.I. Fears

Dario Amodei, Anthropic’s chief executive, on Tuesday.

NYT Technology 18m ago

AI has transformed the Pentagon’s aging networks into a national security risk

For nearly three decades, the Pentagon has put off upgrading its antiquated computer networks, instead focusing its funds on cutting-edge weapons systems. But as artificial intelligence grows in power, it has become increasingly capable of finding low-level weaknesses in the agency’s digital systems and morphing them into debilitating intrusions that could significantly compromise national security. Democracy Dies in Darkness AI has transformed the Pentagon’s aging networks into a national...

Washington Post Tech 18m ago