Several members of the U.S. House of Representatives, led by Rep. Adam Schiff, are calling on Google to explain its AI Overviews feature. This tool, integrated into the Google search engine, summarizes information using AI.
Schiff and other lawmakers wrote to Google CEO Sundar Pichai about the feature’s inaccuracy and danger. The letter further states that the AI has been pulling information from satirical websites and citing long-debunked conspiracy theories, providing unsuspecting users with wrong information.
Letter raises misinformation concerns
Schiff underlined the risks of fake information, particularly in the background of Google as a primary source of news and health information for many Americans. The letter, also signed by Representatives Henry C. Johnson, Jr. (D-Ga. ), Donald S. Beyer Jr. (D-Va. ), Pramila Jayapal (D-Wash. ), and Lori Trahan (D-Mass. ), seeks clarification on several aspects of the AI Overviews feature.
The lawmakers also want to know how users are informed about the risks of misinformation and the time frame for correcting or removing false information.
Google’s AI Overviews feature employs the Gemini AI model to extract information from the top-ranked pages and provide a summary with links. However, the feature has been accused of giving out wrong and misleading information. Some of the examples include advising people to consume rocks for vitamins and endorsing the religion of Barack Obama.
Recent controversies highlight issues with AI-generated content
In May 2024, Google was criticized shortly after the release of AI Overviews in the USA because the feature suggested adding glue to pizza so that cheese would stick. At press time, Google has not issued a direct response to the lawmakers’ letter.
In a blog post, Liz Reid, Google’s Vice President and head of search, also explained that there were problems with the AI Overviews feature. Reid pointed out that one of the issues that AI has is that it does not perform well when the queries asked are not backed up by sufficient high-quality data. She pointed out that when a user enters a specific or rare topic or a topic that contains satire, for instance, the AI will choose answers from the limited sources it has, and sometimes the sources are not credible.
There are also discussions about the financial consequences for publishers. AI-based search engines such as Google may present a summary of the content from different sources without directing traffic back to the originators, which may affect their revenue. This has led to legislative actions such as those in California, where legislators advocate payment to publishers when tech companies utilize their content.
