Live Roundups Editor's Desk Insights Executive Ops Weather About 🔍
Home › Technology › OpenAI, Anthropic Model Tests Reveal More ‘Unsanctioned’ Actions
Technology

OpenAI, Anthropic Model Tests Reveal More ‘Unsanctioned’ Actions

Bloomberg Technology Wednesday 05 August 2026, 11:45 UTC 1 min read
Key Points

Aug 5th, 2026 OpenAI, Anthropic Model Tests Reveal More ‘Unsanctioned’ Actions Artificial intelligence models developed by OpenAI and Anthropic PBC carried out “unsanctioned” actions — including hacking a website and attempting to inject harmful code into software during safety testing — reinforcing fears that neither the creators nor seasoned researchers of these systems can predict their actions in testing. Bloomberg Opinion columnist Parmy Olson joins Francine Lacqua on "The Pulse" to...

Aug 5th, 2026 OpenAI, Anthropic Model Tests Reveal More ‘Unsanctioned’ Actions Artificial intelligence models developed by OpenAI and Anthropic PBC carried out “unsanctioned” actions — including hacking a website and attempting to inject harmful code into software during safety testing — reinforcing fears that neither the creators nor seasoned researchers of these systems can predict their actions in testing. Bloomberg Opinion columnist Parmy Olson joins Francine Lacqua on "The Pulse" to discuss the implications.
PBC (ORG) Bloomberg Opinion (ORG) Parmy Olson (PERSON) Francine Lacqua (PERSON)
Originally published by Bloomberg Technology Read original →

Related Stories

Samsung, SK Hynix pledge stronger shareholder returns

Samsung, SK Hynix pledge stronger shareholder returns Aug 5 : Samsung Electronics is exploring ways to enhance shareholder returns in a "sustainable manner," the South Korean company said in a statement to Reuters on Wednesday. Its smaller rival SK Hynix also said in a separate statement to Reuters that the company is preparing concrete shareholder return plans by the end of the year and expects to "meaningfully expand" capital return. Last week, the two companies reported record-smashing...

Channel News Asia 24m ago

The Winners of Trump’s A.I. Safety Plan

From left, OpenAI’s Sam Altman, President Trump and Demis Hassabis, the C.E.O. of Google DeepMind, at a G7 Summit event in June.

NYT Business 34m ago

Three Six Mafia – Data about "6/6/6 dating" (2024)

Three Six Mafia Most guys my age have "the chat". The one with your college buddies that only exists to share memes, argue, bully, and occasionally announce that you got a promotion or had a kid. Recently the discussion switched to the Three Six Rule.

Hacker News 37m ago

13 Best Coolers for Sunshine and Nighttime (2026)

We've been hunting down the best coolers for years. Whether you're heading out for an evening picnic, a weekend adventure, or a weeklong overlanding trip, you need to chill—your food and drinks, that is. There's a wealth of options these days, from little soft-sided coolers perfect for a day trip to heavy-duty, high-performance wheeled coolers with ice retention times that seem to defy the sun—like our top pick, the Yeti Tundra Haul.

Wired 41m ago