Ars Technica Fires Reporter for AI-Fabricated Quotes in Article
Ars Technica fired senior reporter Benj Edwards after the publication retracted an article containing AI-generated fabricated quotes attributed to engineer Scott Shambaugh. Edwards acknowledged using AI tools while working sick and inadvertently including paraphrased words as direct quotes, which he failed to verify against the original source.
What happened
Ars Technica published an article on February 13, 2026, by Benj Edwards and Kyle Orland covering an incident in which an AI agent published a hit piece about engineer Scott Shambaugh. The article included quotes attributed to Shambaugh that were not his actual words.
Edwards used an experimental Claude Code-based AI tool to extract source material from Shambaugh's blog post, but when the tool failed due to content policy violations, he pasted the text into ChatGPT to understand why. In the process, he inadvertently obtained paraphrased versions of Shambaugh's words rather than his actual quotes, which he then included in the article without verification.
Shambaugh pointed out that he had never said the attributed quotes. Ars Technica retracted the article on February 15, 2026, and editor-in-chief Ken Fisher published an editor's note apologizing for the "fabricated quotations generated by an AI tool and attributed to a source who did not say them."
Reported impact
The retraction represented what Fisher described as "a serious failure of our standards" at an outlet that has covered AI risks for years and had a written policy requiring AI-generated material to be clearly labeled.
Index analysis
Edwards stated he was working from bed with a fever and little sleep while sick with COVID when he made the error. He emphasized that the article's text was human-written, the tool was not being used to generate the article but to help structure references, and the incident was isolated and not representative of Ars Technica's editorial standards.
Governance controls
Ars Technica stated that AI-generated material must be clearly labeled according to company policy, and that none of its articles are AI-generated. Fisher's review of recent work found no additional issues and characterized the error as an isolated incident.
Following the incident, Ars Technica terminated Edwards' employment as a senior AI reporter.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.