Some media outlets are expressing concerns about AI-generated summaries displayed in top Google search results, suggesting that these summaries are reducing news readership. Experts continue to highlight worries about the accuracy of these AI summaries.
The introduction of Google’s AI Overview feature, which gained attention due to notable errors like recommending the use of glue for pizza toppings, raised doubts about the reliability of the technology. Jessica Johnson, a senior fellow at McGill University’s Centre for Media, Technology, and Democracy, emphasized the lack of public discourse surrounding the significant impact of this technological shift on search behavior and daily lives.
Despite users reporting errors in AI-generated summaries, there is no definitive academic research quantifying the extent of inaccuracies. A recent BBC report examining AI chatbots from major companies found “significant inaccuracies” in their news story summaries, although Google AI Overviews were not specifically evaluated.
Google includes a disclaimer in small font at the bottom of its AI summaries, cautioning users that “AI responses may include mistakes.” The company asserts that the accuracy of AI summaries matches other search features, such as featured snippets, and is committed to enhancing response quality and usefulness.
Chirag Shah, a professor at the University of Washington, specializing in AI and online search, explained that the error rate in AI systems stems from their functioning. Generative AI relies on vast datasets for predictions rather than understanding concepts as humans do, leading to potential flaws in retrieved information. Shah highlighted the inherent challenge of ensuring accuracy in AI systems due to their inability to discern nuances like jokes or sarcasm.
As AI integration reshapes search functions, other companies, including OpenAI’s ChatGPT, are deploying generative AI systems as standalone search engines despite known flaws. Shah noted a shift in online platform design focusing on user retention within the same system, potentially altering the nature of the internet and search usage.
A Pew Research Center study revealed that users are less likely to click on links following AI summaries compared to traditional search results, posing revenue concerns for news publishers worldwide. The impact of AI summaries on news consumption has prompted action, with independent publishers in the UK filing complaints regarding significant harm caused by AI overviews.
The evolving landscape of AI and search engines poses challenges for news publishers, with concerns raised about revenue loss and reduced visibility for content. The future of online news legislation remains uncertain, with potential implications for the media industry’s sustainability amid technological changes.
