last week, One of my WIRED articles was used in an unexpected way in Google's AI summary search results, making me worry about the future of journalism.
I was trying out AI Outline, the company's new generative AI feature designed to answer online queries. Having asked multiple questions about topics I've covered recently, I wasn't surprised to see my article linked as a footnote at the bottom of the box containing the answer to my query. I was surprised, however, by how much of the AI Outline's first paragraph was a direct quote from my writing.
Below, the screenshot on the left is from an interview with one of Anthropic's product developers about tips for using their Claude chatbot: The screenshot on the right is from a part of Google's AI overview that answered questions about using Anthropic's chatbot: Read the two paragraphs side by side, they remind me of a student cheating in class who copies the answers to my homework and barely bothers to rephrase them.
Reese Rogers (via Google)
Without AI Summary enabled, my articles would often be featured snippets highlighted at the top of Google search results, providing a clear link for curious users to click when they were looking for advice on using the Claude chatbot. During early tests of Google's new search experience, featured snippets for articles continued to appear for related queries, but they were pushed below AI Summary answers that pulled from my reports and inserted parts of them into a 10-item bullet-point list.
In email correspondence and phone calls, Google spokespeople acknowledged that AI-generated summaries may use some text taken directly from web pages, but defended the AI summaries as clearly referencing the original sources. In my case, the first paragraph of the response was not directly attributed to me; instead, my original article was one of six footnotes hyperlinked near the bottom of the result. With the source link that far down, it's hard to imagine a publisher getting significant traffic in this situation.
“AI Overviews conceptually match the information that appears in the top web results, including those linked to in the Overview,” a Google spokesperson said in a statement to WIRED. “This information is not a replacement for web content, but is designed to help people get a feel for what's out there and click to learn more.” Looking at the word choice and overall structure of the AI Overview in question, I disagree with Google's characterization that the results are merely a “conceptual match” of my writing. It's much more than that. And even if Google's developers never intended for the feature to replace the original work, AI Overviews provide a direct answer to the question, burying citations and reducing the incentive for users to click through to the source material.
“We've found that links included in AI Overviews receive more clicks than if the pages were surfaced as traditional web listings for that query,” a Google spokesperson said. WIRED wasn't provided with data to back up this claim, so it's impossible to independently verify the AI feature's impact on click-through rates. It's also worth noting that the company compared AI Overview referral traffic to traditional blue link traffic from Google, not to articles that were selected for featured snippets, which may have much higher click-through rates.
While many of the AI cases remain unresolved, one copyright lawyer I spoke to was skeptical that I would win if I did sue. “I don't think it's a strong case for copyright infringement,” says Janet Freese, an attorney at Faegre Drinker Biddle & Reath. “Copyright law is generally concerned with not interfering with useful or helpful things.” Her perspective focused on the type of content in this particular example of original work, explaining that instructional or factual writing, like my advice column, versus a more creative piece like poetry, would be much harder to argue.
I’m certainly not the first to recommend focusing on your target audience when writing prompts for chatbots, so I agree that the factual aspects of my writing complicate the overall situation, but it’s hard to imagine a world in which Google would land on that very paragraph about Claude’s chatbot in their AI Overview results without referencing my work.
