Major Newsrooms Announce New AI Reporting Policies as Industry Scrambles to Set Standards

18 September 2026

Associated Press has revised its standards for AI in the newsroom, as more news organizations formalize rules for using AI to assist reporting. The updated guidance, announced September 18, allows AI to aid journalists in specific tasks such as research, document summarization, transcription and translation—with strict requirements that human journalists must review and edit all AI-generated output before publication.

The AP's new rules add to a wider wave of newsroom policy-making on AI as the technology becomes more woven into journalism. An analysis of AI guidelines at 52 news organizations, published in July 2023, found that newsrooms from USA Today to the Financial Times had begun to delineate how they would use and govern AI, what limitations they would place on it, and what their disclosure practices would be.

The AP guidelines are more permissive than its earlier position, which said in 2023 that the service would not publish AI-generated content or images. In a policy update published in 2026, however, AP said it would allow AI to assist with "early-stage research, document summarization, transcription, translation, headlines, story summaries, shotlists, and grammar, spelling, and search optimization."

Newsrooms Boost Human Oversight of AI

When deploying AI in the newsroom, many organizations are placing strong limits on the technology. Reuters' guidance, for instance, said the outlet would not create text using generative AI. AP's new standard includes similar limits, but stretches the scope slightly to encompass metadata and ancillary content—with disclosure requirements should AI be used.

"The use of generative AI to create text, video or audio that could be considered journalism, not just metadata, is likely to remain intensely controversial," the Columbia Journalism Review wrote in 2023, as news organizations including AP began to set ground rules.

"The CJR stresses human oversight, transparency, accountability, and authenticity in newsroom AI policies," the CJR report said.

Like other news organizations, the AP requirements specify that while AI can assist journalists, news judgment remains a human responsibility and AI output must be reviewed. Automation is allowed as a journalist aid, not as a replacement for human reporting and editing.

By mandating that "all AI-generated output be reviewed and edited by AP journalists before publication," AP has added structure to its staff's use of AI. Those rules will apply to summaries, translations, and headline-suggesting AI. The AP guidelines also introduce new standards for "reporting on AI-generated and manipulated content" and for AI coding assistants in software development, combining its old rules into the updated standard.

European Newsrooms Comply with EU AI Act

European news organizations are facing a new deadline for transparency in AI-generated content as Article 50 of the EU's Artificial Intelligence Act came into effect on August 2, 2026. The Act requires entities in the EU to disclose if they publish AI-generated text on matters of public interest. The disclosure requirement does not apply to text that has been edited under editorial responsibility.

An analysis by the independent observatory iMEdD says in many cases, editorial review and control should serve as a sufficient safeguard for newsrooms even without formal disclosure, as most newsroom AI guidelines rely on human oversight.

iMEdD also reported the emerging consensus is that AI has a bright future as a journalist's assistance tool for researching, summarizing, language translation, and some areas of content generation—but journalists and editorial boards maintain control over what content makes it to the newsfeed, chatbot, or archive.

In newsrooms where staff use research and translation AI, iMEdD noted, human editing is often already default practice and may provide more useful oversight than a disclosure line. The role of human editing "can prevent errors or misinformation from being amplified, whether unintentionally by AI malfunction, or intentionally by bad actors using generative AI."

AI and Newsroom Policy in 2026

A survey of 52 news organizations found that about 71% of newsroom AI policies mentioned journalistic values; about 69% mentioned AI pitfalls; and about 90% required disclosure if AI was used in a story or investigation. Today's standardized approaches to AI governance reflect both newsroom experiments with the technology that predated disclosure law—along with the contours of the EU's AI Act.

With the new AP guidelines and those already in use by news organizations across the English-speaking world, newsrooms are sending a clear message in 2026: AI can help, but human judgment rules. As the AP puts it, the human newsroom standard remains: "verification, accountability and editorial judgment."