When will my queries trigger a summary of AI-generated results? “We’ll show you an AI summary for complex queries,” Google spokesperson Mallory de Leon said. Masu. “You’ll see an AI overview in Google search results when our systems think generative AI is particularly useful, such as when you want to quickly understand information from a variety of sources.” First In our tests, it felt like the AI summaries were displayed almost randomly for queries, and the summaries appeared for simple questions as well as more complex questions.
De Leon said AI Overview leverages a customized version of Google’s Gemini model, adding elements from the company’s search system, including a knowledge graph containing billions of common facts.
One of my main hesitations about rolling out this feature is the potential for continued AI illusions, more commonly known as lies. When you interact with Google’s Gemini chatbot, you’ll see a disclaimer at the bottom that says, “Gemini may display inaccurate information, including information about people, so please double-check your responses.” . No such disclaimer is added at the bottom of the AI overview, which often simply says, “Generative AI is experimental.”
When asked why he didn’t mention the potential hallucinations of AI Overview, de Leon emphasized that Google still wants to provide high-quality search results, and that the company has We mentioned that we conducted adversarial red team testing to uncover key weaknesses.
“This generative AI implementation is rooted in search’s core quality and safety systems, with built-in guardrails to prevent low-quality or harmful information from surfacing,” she says. “AI Overview is designed to highlight information that can be easily verified by the supporting information we uncover.”
Even though you know this, you may still want to double-check that the information is actually correct by clicking on the links on the web page. However, it’s unlikely that many users, who often want answers right away, would spend the extra time reading the source material for answers generated by Google’s AI.
Google’s head of search, Liz Reid, recently told my colleague Lauren Good that AI Overview will soon reach 1 billion people, as it will be made available to countries other than the United States by the end of 2024. More people said they may come across this new feature. As someone whose job it is to actually spend time with readers clicking on links and reading articles, I am of course concerned about this change. And I’m not alone.
Beyond publisher concerns, it remains unclear what additional impact Google’s AI overview may have on users. Sure, OpenAI’s ChatGPT and other AI tools are hugely popular in the Silicon Valley tech world, but this feature also opens the door to AI-generated chatbots for billions of people who have never used a chatbot before. You will be exposed to a variety of texts. Although AI summaries are designed to save time, they can lead to unreliable results.
