AI agent went rogue and hacked startup by…

Global Security · Direct source

AI agent went rogue and hacked startup by itself, OpenAI reveals

Firm behind ChatGPT reveals agent powered by its models chose to attack Hugging Face database by itself OpenAI has revealed an autonomous AI agent powered by its technology went rogue during a test, accessed the open web and hacked a prominent startup by itself in an “unprecedented incident”. The company behind ChatGPT said the startup Hugging Face had detected and contained the agent – an AI tool designed to carry out tasks without human assistance – which had entered its systems. Continue reading...

Source published: 22 Jul 2026, 10:37 CEST
Publication time from the source RSS/feed
Captured by GC: 22 Jul 2026, 12:55 CEST
When GlobalsConflicts first captured this item.
Source: The Guardian - World
Trust: medium · direct source trail
Source quality: usable
direct source; further independent sources matter for hard confidence
Actors: United States, UN, Lebanon, Houthis, OCHA, Yemen

Brief

medium direct source trail The evidence trail is rated, not absolute truth.

What is reported

Visible evidence

Source published (RSS): 22 Jul 2026, 10:37 CEST. Publication time from the source RSS/feed
The report is assigned to the Global Security dossier.
The visible source is The Guardian - World.

Still unclear

5 direct reports nearby, but not automatically the same core claim.
5 related reports in the same dossier may add context.
The page rates the evidence trail, not the political truth of a position.

Why it matters

This report is assigned to the Global Security dossier. It matters because it adds a concrete new trail in the current source window. The brief uses 6 sources in the surrounding context while keeping timestamp, publisher and original URL visible.

Trust assessment

Direct source with related reports nearby. The evidence trail is usable, but should not be read as a fully confirmed situation yet.

Editorial boundary

Still open: whether further independent sources confirm, correct or merely repeat the same development. The trust level describes the source trail, not absolute truth.