Ars Technica publishes its newsroom AI policy: no AI-generated sources, no synthetic documentation

Ars Technica has formalized and published the AI rules that have governed its newsroom since generative tools became available. The core prohibition: AI cannot be used to generate, extract, or summarize material that ends up attributed to a named source, whether as quote, paraphrase, or characterization. Claims cannot rest solely on AI-generated summaries, and reporters cannot describe material as ‘reviewed’ unless they examined it themselves. Any use of AI in reporting must be disclosed to editors, with the human author retaining full responsibility for the output.

On visual content, the policy bars AI-generated imagery, audio, or video from being presented as authentic documentation of real events, and prohibits alterations that change the meaning of documentary media. Routine production work like cropping and color correction remains acceptable. When synthetic media appears in coverage of AI itself, it must be clearly labeled as AI-generated, with the disclosure placed adjacent to the material.

The accountability clause is the part worth noting: responsibility for accuracy cannot be offloaded to editors, colleagues, or the tools themselves. Ars frames the publication of these rules as a transparency move — readers get to see the standards rather than take them on faith. The policy was last updated April 22, 2026.