Google's AI Overviews Are There Whether You Want Them or Not - Latest Global News

Google’s AI Overviews Are There Whether You Want Them or Not

When can you expect your query to trigger an AI-generated summary of results? “AI overviews are displayed for complex queries,” says Mallory De Leon, a Google spokeswoman. “AI overviews appear in your Google search results when our systems determine that generative AI can be particularly useful – for example, when you want to quickly understand information from different sources.” During my initial testing, I had the feeling that the AI summaries appeared almost randomly during queries and the summaries appeared for both simple questions and more complicated questions.

According to De Leon, the AI ​​overview is based on a customized version of Google’s Gemini model, supplemented by aspects of the company’s search system, such as the Knowledge Graph, which contains billions of general facts.

For some AI overview answers, the website links are immediately visible. For other AI overviews you must click show more to see where the information comes from.

Google Reece Rogers

One of my main concerns about this feature when it launches is the ongoing potential for AI hallucinations, more commonly known as lying. If you interact with Google’s Gemini chatbot, a disclaimer below reads: “Gemini may display inaccurate information, including about people. So please double-check the answers.” There is no such disclaimer at the end of the AI ​​overview, which often simply reads: “Generative AI is experimental.”

When asked why potential hallucinations for AI Overviews are not mentioned, De Leon emphasizes that Google wants to continue offering high-quality search results and mentions that the company has conducted controversial red-teaming tests to uncover potential vulnerabilities in the feature.

“This implementation of generative AI is built on Search’s core quality and safety systems and has built-in guardrails to prevent low-quality or harmful information from surfacing,” she says. “AI Overviews are designed to highlight information that can be easily verified based on the supporting information we display.”

Knowing this, you may still want to click through the website links to double-check that the information is actually correct. Although it is hard to imagine that many users who are often looking for quick answers spend extra time reading the source material for Google’s generated answer.

Liz Reid, head of Google Search, recently told my colleague Lauren Goode that AI Digests are expected to be available to countries outside the US before the end of 2024, meaning over a billion people will likely soon be exposed to this new feature . As someone whose job depends on readers actually clicking on links and taking the time to read articles, I’m naturally concerned about this change – and I’m not alone.

Aside from publishers’ concerns, it also remains unclear what additional impact Google’s AI overviews might have on users. Yes, OpenAI’s ChatGPT and other AI tools are very popular in Silicon Valley tech circles, but this feature will likely introduce AI-generated text to billions of people who have never used a chatbot before. Even though AI overviews are designed to save you time, they can produce less trustworthy results.

Sharing Is Caring:

Leave a Comment