Google’s parent company lost more than $70 billion in market value in a single trading day after its “woke” chatbot’s bizarre image debacle stoked renewed fears among investors about its heavily promoted AI tool.
Shares of Alphabet sank 4.4% to close at $138.75 in the week’s first day of trading on Monday. The Google’s parent’s stock moved slightly higher in premarket trading on Tuesday.
The massive hit came after Google was forced to pause its Gemini chatbot’s image creation tool – which was churning out factually and historically inaccurate images such as Black Vikings, female NHL players, “diverse” versions of America’s Founding Fathers and even an Asian woman dressed in a Nazi-era German soldier’s uniform.
The chatbot’s bizarre behavior could fuel public concerns that Google is “an unreliable source for AI,” Melius Research analyst Ben Reitzes warned in a note to investors.
“We have been arguing that Search behavior is about to change – with new AI-infused features,” Reitzes said in the note, according to Bloomberg. “This ‘once in a generation’ change by itself creates opportunities for competitors, but even more if a meaningful portion of users grow concerned about Google’s hallucinations and bias.”
The stock plunge also came on the same day that Google DeepMind CEO Demis Hassabis, one of the company’s top AI bosses, admitted that Gemini’s image tool would be offline for a “few weeks” while the issue was fixed.
“We have taken the feature offline while we fix that. We are hoping to have that back online very shortly in the next couple of weeks, few weeks,” Google DeepMind CEO Demis Hassabis said Monday.
Hassabis added that the tool was not “working the way we intended,” during his appearance at the Mobile World Congress in Barcelona.
Critics have pointed to Google Gemini’s bizarre behavior as evidence that it received politically biased “training” from the company’s AI experts.
Those complaints surged last week after politically charged tweets by Google Gemini’s product lead Jack Krawczyk allegedly resurfaced last week. In the old posts, which went viral on X, Krawczyk allegedly stated that “white privilege is f—king real” and America is rife with “egregious racism.”
Since the company does not publicly reveal details about the parameters that govern the chatbot’s behavior, it’s difficult to get a clear explanation for why it was generating the so-called “woke” images.
Additional testing of Google Gemini’s responses raised further alarms – with the chatbot refusing to condemn pedophilia and claiming there was “no right or wrong answer” when asked if Adolf Hitler or Elon Musk is worse.
