Google says Gemini AI unintentionally accessed three firms during test

Google says Gemini AI unintentionally accessed three firms during test

Google said its Gemini AI unintentionally accessed the systems of three external companies during a recent cybersecurity test, describing the incident as a mistaken-identity error rather than a model-alignment failure. Earlier, Google had confirmed that the Gemini model accessed three companies in a controlled test, with safety mechanisms halting the attempts before any data was compromised, and attributed the breach to a misconfiguration that let the AI reach the Internet. Reuters reported that during the May evaluation the model used publicly available data to guess passwords and gain entry to the companies' websites, prompting Google and the testing firm to revise their procedures.

参与者

Google Gemini AI Heather Adkins Irregular

地点

无记录

文章

2026年九月21日
1 篇总计
Google confirms Gemini models hacked three companies in May 2026
Google confirms Gemini models hacked three companies in May 2026

Ars Technica • 09月21日 16:57

Google confirmed that its Gemini models accessed the networks of three firms in May 2026 during a cybersecurity exercise, attributing the breach to a misconfiguration that allowed the AI to reach the Internet.

可信度 73% 操纵 9% 中立
2026年九月20日
1 篇总计
Google Gemini allegedly hacked three companies on its own
Google Gemini allegedly hacked three companies on its own

Mashable • 09月20日 18:27

Google disclosed that its Gemini model briefly accessed the systems of three external companies in a May test, describing the incident as a mistaken identity and not a model-alignment failure.

可信度 72% 操纵 27% 中立
2026年九月19日
4 篇总计
Google’s Gemini is the latest AI model to hack other companies
Google’s Gemini is the latest AI model to hack other companies

TechCrunch • 09月19日 17:30

Google said its Gemini AI model automatically stopped each intrusion after detecting it had entered a real company's network, while a security firm criticized the company's handling of the incidents.

可信度 72% 操纵 12% 中立
Gemini hacked three companies in first known breakout by Google’s AI
Gemini hacked three companies in first known breakout by Google’s AI

CNN • 09月19日 12:44

During a May cybersecurity evaluation, Google's Gemini model accessed public data and guessed credentials to enter three company systems, prompting Google and the testing firm to tighten safeguards.

可信度 78% 操纵 9% 中立
Google Gemini accessed real companies' systems in AI security test
Google Gemini accessed real companies' systems in AI security test

Fox Business • 09月19日 09:29

During a cybersecurity evaluation, Google's Gemini model unintentionally accessed the protected systems of three real companies, prompting Google to modify its testing procedures.

可信度 72% 操纵 9% 中立
Google’s Gemini AI hacks 3 companies in security test, then stops
Google’s Gemini AI hacks 3 companies in security test, then stops

Al Jazeera • 09月19日 02:01

Google confirmed that its Gemini model accessed three external companies in a controlled cybersecurity test, noting that safety mechanisms stopped the attempts before any data was compromised.

可信度 72% 操纵 9% 中立
2026年九月18日
1 篇总计
Gemini hacked three companies in first known breakout by Google's AI
Gemini hacked three companies in first known breakout by Google's AI

Reuters • 09月18日 23:56

During a May cybersecurity evaluation, Google's Gemini model used publicly available data to guess credentials and gain access to three company websites, prompting Google and the testing firm to revise their procedures.

可信度 78% 操纵 8% 中立