TechFeb 28, 2024
Alphabet’s New AI Model Was Supposed to be a Gamechanger. Instead, It’s Controversial, Biased, and Problematic.
Search Giant
Noah Weidner
Discover
Alphabet, Google’s parent company, has a problem with its AI. No, it’s not stalking people’s search history, tracking your every move, or responding to the company’s growing stack of antitrust lawsuits… yet. Actually, its problem is that it’s not really that intelligent at all.
Last year, when Google rolled out its first mainstream AI model, Bard, employees begged not to release it — calling it “worse than useless” and “a pathological liar” while complaining of ethics concerns. The company launched the product anyway — but a year later, Bard’s ghost is back to haunt the company.
From hype to hate: The company’s new AI model, Gemini, was intended to position Alphabet as a serious contender to ChatGPT. While the model initially excited users, it’s now frustrating them with “offensive” text responses and imagery.
Alphabet CEO Sundar Pichai called the problematic responses “unacceptable” and urged the AI team to work around the clock to resolve its issues. It’s the first time its CEO publicly acknowledged that Google has a problem building relevant, useful, and long-lasting products.
Where’s the vision? Google’s AI controversy adds to charges made by former employees and founders that Google lacks leadership and is inefficient. Pichai has shuttered many of the company’s 293 products and laid off over 13K employees in a Meta-esque push for efficiency. But where Meta and Alphabet differ are their returns, with Meta (NASDAQ:META) up nearly 40% this year while is down 2%. This places Alphabet alongside Tesla (NASDAQ:TSLA) as one of the only two Magnificent 7 stocks experiencing a decline in 2024.
