A look at Smart Answers, an AI chatbot trained only on Macworld, PCWorld, Tech Advisor, and TechHive content, and offered to the outlets' readers on August 1
By and large, the goal has been: can we make more pages for ads without paying more writers? — Now, a group of tech outlets … Threads: @dsilverman . X: @mr_existant and @harrymccracken Threads: Dwight Silverman / @dsilverman : This is fascinating, and may be a best-use-case for an LLM in news. https://www.theverge.com/... Though what kind of hallucinations will this generate? X: @mr_existant : SEO over the years has degraded to a point that many top webpages on the first page provide little of what you're looking for. https://www.theverge.com/... @harrymccracken : My old stomping grounds created its own AI search feature. (It got one of the first questions I asked—"What is the current version of Windows?"—wrong.) https://www.theverge.com/...
Context & Ripple Effects
Smart Answers turns a set of publishers’ existing editorial archives into a reader-facing answer interface, rather than relying on a general-purpose model’s broader web knowledge. It arrives as AI search products were already demonstrating cited, conversational answers, including Bing’s citation-backed answer experience.
The commercial logic is explicit: the outlets are seeking additional ad-supported pages without adding writers. That places the product alongside experiments to bring advertising into chat responses, such as Microsoft’s planned paid links and industry ads in Bing chat.
First-order effects
- Macworld, PCWorld, Tech Advisor, and TechHive gain a new way to package their own archives into answers and potentially generate more ad inventory without commissioning equivalent new articles.
- Readers receive answers constrained to those outlets’ material, but the reported incorrect response about the current Windows version makes accuracy and trust an immediate product risk.
Second-order effects
- Editorial teams and publishers adopting similar tools will need to weigh lower-cost answer production against the review, correction, and reputation costs created by hallucinated or stale responses.
- If answer pages become a meaningful ad surface, publishers will face pressure to define how sponsored links or ads are distinguished from editorially sourced answers.
Third-order effects
- The experiment points toward publishers treating proprietary archives as inputs for owned answer engines, shifting some value from page-by-page search discovery to direct conversational retrieval.
- Whether that model endures depends on whether publishers can preserve attribution, accuracy, and reader trust while monetizing answers that may reduce the need to browse underlying articles.
The trend: Publishers are testing controlled AI answer layers over their own content as they seek to retain audience and advertising value in a conversational-search market.