Google’s new AI Overview search feature has sparked controversy as it appears to directly replicate content from existing articles without adequate source attribution. I recently found one of my articles published on WIRED being utilized in an unforeseen way by Google’s AI Overviews – a tool designed to generate answers to online queries.

For context, I’d used the AI Overview feature and was not surprised to see my piece footnoted at the end of the response box, given the topics I queried were those I’ve recently written about. However, what took me by surprise was the similarity between my original piece and the generated AI Overview.

I compared the AI Overview to an interview I’d conducted with a product developer at Anthropic, specifically about utilizing the company’s Claude chatbot. The degree of resemblance between the two – to the point of feeling plagiarised – was remarkable. My original piece was often a featured snippet on the top of Google’s search results, offering anyone seeking advice on the Claude chatbot a direct link to my article. However, with AI Overviews switched on, my piece, though still featured, was pushed below the AI-generated answers which appeared to incorporate elements of my work in a bullet-point list.

Google defended these AI summaries, claiming they highlight original sources, however, the result used phrases and ideas from my work without direct attribution within the text. While Google asserts that AI Overviews are not replacing original content, they do offer direct answers that bury attributions and discourage users from clicking through to original sources.

Google’s claim that AI Overviews yield higher clicks than traditional website listings remains I unverified, as they provided no supporting data. The comparison made was also flawed as it did not account for featured snippets, which likely have higher click-through rates.

This feature’s potential negative impact on digital journalism cannot be overstated. It could considerably alter the nature of internet searches, the value of featured snippets, and challenge the sustainability of digital journalism. This new development underscores the precarious relationship between Google and online publishers and exemplifies how changes to Google’s search service can drastically impact traffic and potentially threaten the existence of entire publications.

The precarious state of the journalism industry is driving publishers to seek licensing deals with major AI companies. The Associated Press, The Atlantic, the parent company of Business Insider and Politico, and other significant players all have contracts with OpenAI. This trend underscores the value of reputable journalism to AI companies even if they circumvent the licensing of the original content.

Google AI’s reliance on user-generated content is not without flaws. Google’s head of search, Liz Reid, recently blamed discussion forums for some spectacularly inaccurate AI Overview responses. These fallacies serve as a reminder of the critical importance of fact-checking and information validity if direct AI-summarised responses are the way forward. It also underlines that, if you’re going to copy a homework, at least make sure it’s from an A-grade student, rather than a random commenter on Reddit.