OpenAI shows how newsrooms around the world use its AI in journalism

OpenAI has published an overview on its website of how news organizations around the world use its AI tools to strengthen reporting, grow audiences and improve business operations.

Three tasks AI handles

OpenAI identifies three areas where newsrooms apply its technology. By the company's own account, AI helps journalists and publishers work more efficiently without replacing journalism itself.

• Reporting — processing large document sets, transcribing interviews, finding facts and preparing drafts.
• Audience — personalizing how content is presented, working on headlines and formats, attracting new readers.
• Business operations — automating routine processes inside a publishing house.

The announcement does not explicitly list which tools are involved. But the products in question are OpenAI's own: ChatGPT and the API, on which newsrooms build their own workflows.

Why newsrooms want it

Newsrooms turn to AI to take routine work off journalists' plates and free up time for reporting. Typical scenarios where neural networks are already used in newsrooms worldwide include transcribing audio and video, translating material, summarizing long documents and performing initial data analysis.

For example, the Associated Press has been automating financial and sports briefs from templates since 2014, making it one of the earliest examples of AI in a newsroom.

For media companies, AI is both a way to speed up content production and a growth tool: personalized feeds keep readers engaged longer. At the same time, newsrooms keep humans in control. Because of the risk of errors and "hallucinations," final fact-checking stays with journalists, and many major outlets have already published their own editorial policies on the use of neural networks.

How newsrooms use AI in practice

Some examples are already live in public. In 2024, The Washington Post launched a chatbot, Ask The Post AI, that answers readers' questions based on the paper's reporting. The Financial Times is testing an assistant, Ask FT, built on its archive. Agencies such as Reuters and Bloomberg have used algorithms for years to automatically generate financial summaries and process data.

These cases point in the same direction: AI takes over routine information processing, while journalists concentrate on verification, analysis and exclusive stories.

What figures and deals did OpenAI cite?

OpenAI's brief announcement gives no concrete metrics: no number of participating publications, no estimate of time saved, no names of individual partners. The piece is structured as a thematic overview of focus areas, not as a report with numbers.

"News organizations use AI to strengthen reporting, grow their audiences and improve business operations,"

OpenAI's blog says.

The publication fits OpenAI's broader strategy of getting closer to the media industry. The company signed its first licensing agreement, with the Associated Press, in July 2023. Deals followed with Axel Springer (December 2023), the Financial Times (April 2024) and News Corp (May 2024, in a deal the press valued at more than 250 million dollars over five years), as well as with Vox Media, The Atlantic and Condé Nast. At the same time, OpenAI remains a defendant in the lawsuit filed by The New York Times in December 2023, in which the paper accuses the company of training models on its articles without permission.

What this means

AI in newsrooms is no longer an experiment. It is becoming part of the everyday production cycle, from interview transcription to feed personalization. For OpenAI, the overview is also a way to gain a foothold in the media industry, a relationship that has remained contradictory since 2023: licensing deals coexist with lawsuits. For readers, this means faster, more personalized news, but also new questions about where the line falls between the work of a human and the work of an algorithm.