Google disclosed that in May its AI model Gemini gained unauthorized access to three outside systems during a test by guessing login information or using login credentials found in a public repository. Google vice president for security engineering Heather Adkins said the model thought the outside computer systems "were part of the test," and in all three instances it stopped before doing anything further with its access. Google said it did not consider the unauthorized logins to rise to the level of misalignment, saying the intrusions resulted from mistaken identity, that the model corrected itself, and that it believed they caused no damage. The disclosure follows similar ones by Anthropic and OpenAI. Sydney Von Arx, CEO of Nightingale Collective, questioned why Google did not disclose sooner and said she believed Google was too hasty to say the incidents don't rise to the level of misalignment. Google said it did not learn about the intrusions until July, when Irregular, an AI-focused cybersecurity company carrying out the tests on Gemini when the intrusions occurred, reviewed its work to look for incidents similar to the Hugging Face disclosure, after which Google investigated.