Google is updating its ubiquitous search engine with the next generation of its artificial intelligence technology as part of an effort to provide instant expertise amid intensifying competition from smaller competitors.
The company announced Wednesday that it will feed its Gemini 2.0 AI model into its search engine so it can field more complex questions involving subjects such as computer coding and math.
As has been the case since last May, the AI-generated overviews will be placed above the traditional web links that have become the lifeblood of online publishers dependent on traffic referrals from Google's dominant search engine.
Google is broadening the audience for AI overviews in the U.S. by making them available to teenage searchers without requiring them to go through a special sign-in process to see them.
The stage is also being set for what could turn out to be one of the most dramatic changes to the search engine's interface since Google co-founders Larry Page and Sergey Brin started the company in a Silicon Valley garage during the late 1990s.
Google is going to begin a gradual rollout of an “AI mode” option that will result in the search engine generating even more AI overviews. When search is in AI mode, Google is warning the overviews are likely to become more conversational and sometimes head down online corridors that result in falsehoods that the tech industry euphemistically calls “hallucinations.”
“As with any early-stage AI product, we won’t always get it right,” Google product vice president Robby Stein wrote in a blog post that also acknowledged the possibility “that some responses may unintentionally appear to take on a persona or reflect a particular opinion.”
Which phones are getting One UI 7 first and when
More stringent guardrails are supposed to be in place to prevent AI mode from steering people in the wrong direction for queries involving health and finance.
The need for additional fine tuning is one reason Google is initially only offering AI mode in its experimental Labs section, and only subscribers to its $20-per-month Google One AI Premium will be allowed to test it out at first.
But these tests almost always result in the technology being released to all comers — a goal that Google is pursuing in response to AI-powered search engines from ChatGPT and Perplexity.
Google's amped-up usage of more sophisticated AI overviews is likely to amplify worries that the summaries will make web surfers even less likely to click on links to take them to sites with useful information on the topic.
Those traffic referrals are one of the main ways that online publishers attract the clicks needed to sell the digital ads that help finance their operations.
Google executives insist AI overviews are still driving traffic to other sites by driving up people's curiosity so they engage in more queries to learn more, resulting in more clicks to other publishers.
But those reassurances haven't placated publishers who believe that Google will be the main beneficiary of AI overviews, further enriching an internet empire that already generates more than $260 billion in annual ad revenue.
The expanded use of AI overviews also could expose Google to more allegations that it is abusing the power of a search engine that a federal judge last year found to be an illegal monopoly in attempt to maintain its position as the internet's main gateway.
The U.S. Justice Department, which filed the monopoly claims against Google in 2020, is now proposing a partial breakup of the company that would include the sale of its Chrome browser as part of its punishment. The hearings on the proposed penalties against Google, which may include digging deeper into its use of AI, are scheduled to begin next month in Washington D.C.
Online educational online service Chegg already has amplified on that monopoly case with a lawsuit filed last month in the same Washington court accusing Google of improperly cribbing information from its site to present in its AI Overviews. Google has denied the allegations.