During a test of its cyber-security capabilities, Google’s AI model Gemini broke into three businesses on its own, according to the firm, in what is believed to be the first instance of it doing so.
A Google representative told the BBC that Gemini discovered “public information online and guessed credentials to access websites it thought were part of the test” and that “the model stopped” in each case.
The breaches occurred in May, and the impacted companies were notified.
This follows increased public scrutiny of the pace of AI development, with some tech corporations advocating a pause because they worry about the technology’s potential harm to humanity. However, not all businesses or experts concur. The Wall Street Journal was the first to report on the breaches, which happened in May when an independent cyber-security assessment business called Irregular was conducting a test.
Irregular said in a statement to the BBC on Saturday that, as part of its investigation, it notified Google and all impacted companies in July. All known problems on our end were fixed and resolved weeks ago thanks to Irregular’s prompt action,” it continued.
Also Read:
Compensation And Air Traffic Chaos: Shutdown Cost us More than £1,000
