Story preview

OpenAI says its models engaged with US government websites in new model misbehavior disclosure

The Washington Post

About this story

Published: Sep 26, 2026, 6:46 AM UTC

How other outlets describe it

OpenAI's decision to hold back the model, called GPT-6.1 Astra, comes amid a broader push within the industry to slow the development of increasingly autonomous systems until safety measures catch up.

NPRUnited StatesOpenAI delays latest model over security concerns, as industry faces pressure

If you feel like we're repeating the ominous beginning of a post-apocalyptic movie, you're officially not crazy. Now, what do we do about it?

AOL.comOpenAI Bots Go Rogue Again, This Time On US Government Sites

Topics in this story

#openai #models #engaged #government #websites

Continue reading on the publisher site for the full article.

Read on publisher →The Washington Post

News Atlas is free and independent. If it’s useful to you, you can support it with a coffee. ☕ Buy me a coffee

At a glance

Outlets covering it

in the last 3 days

First reported

5 days ago

12 outlets since

Countries

Coverage

Tone

Who else is covering this

12 other outlets reported this story in the last 3 days.

OpenAI's agents obscured hacking activity in government site breachesYahoo Finance5 hr. ago
OpenAI says its models engaged with US government websites in new model misbehavior disclosureKREM3 days ago
OpenAI delays latest model over security concerns, as industry faces pressureNPRUnited States2 days ago
OpenAI Ignored Employees’ Warnings About Safely Testing A.I. ModelsThe New York TimesUnited States45 hr. ago
OpenAI Bots Go Rogue Again, This Time On US Government SitesAOL.com2 days ago
OpenAI scraps release of newest AI model over safety concernsAnadolu AjansıTürkiye2 days ago
Anthropic says its AI models pose ‘existential risk to humanity’ in leaked IPO filing: reportWQOW45 hr. ago
OpenAI cancels release of newest model due to safety concernsMacau Business2 days ago
OpenAI scraps launch of latest model over security concernsCP242 days ago
Rogue AI agents tried and failed to hack US and Canadian government websitesAOL.comUnited Kingdom3 hr. ago
OpenAI scraps release of new model due to safety concernsRTE.ieIreland2 days ago
OpenAI pledges task force after hacking of Australia government websiteBusinessLine2 days ago

Who else is covering this

Where in the world it’s covered

Coverage over time

Tone of the coverage

About the publisher

The Washington Post

Visit The Washington Post ↗

Latest from The Washington Post

Home pageGoogle News2,483 days ago
- Google NewsGoogle News7 hr. ago
Google NewsGoogle News6 hr. ago
- Google NewsGoogle News20 hr. ago
Home pageGoogle News2,308 days ago

OpenAI says its models engaged with US government websites in new model misbehavior disclosure

About this story

How other outlets describe it

At a glance

Who else is covering this

Who else is covering this

Where in the world it’s covered

Coverage over time

Tone of the coverage

About the publisher

Latest from The Washington Post

About the publisher

Related topics

About the publisher

Latest from The Washington Post

Who else is covering this

How other outlets describe it