The New York Times has approved artificial intelligence tools that newsroom staff can use to edit stories, summarize information, program and write. According to Semafor, the publication announced in an internal email that product and editorial staff will receive training in artificial intelligence technology and launched a new internal artificial intelligence tool called Echo for summarizing articles, briefings and other company activities.
The company reportedly sent new editorial guidelines to employees detailing the permitted uses of Echo and other artificial intelligence tools, encouraging newsroom staff to use the tools to suggest edits and revisions to their work and to generate summaries, social media copy and SEO headlines.
Other examples mentioned in mandatory training videos shared with employees include using AI to develop news Q&A, quote cards and FAQs, or to advise reporters on what questions they should ask a startup CEO when interviewing him. There are some limitations, though - the company told editors that AI should not be used to draft or substantially revise articles, circumvent paywalls, import third-party copyrighted material, or publish AI-generated images or videos without clear attribution.
It’s unclear how much copy the New York Times will allow to use in published articles. The outlet pledged in a memo issued last year that "The New York Times' journalism will always be reported, written and edited by our expert journalists," and months later it reiterated its commitment to human involvement.
"In May 2024, The New York Times adopted principles for generative AI. We are always accountable for our reporting, no matter how it is created. Any use of generative AI in the newsroom must be based on factual information vetted by our reporters and, like everything else we produce, must be reviewed by our editors."
In addition to Echo, other AI tools that The New York Times has apparently approved for use include GitHub Copilot as a programming assistant, Google Vertex AI for product development, NotebookLM, The New York Times’ ChatExplorer, OpenAI’s non-ChatGPT API, and some of Amazon’s AI products.
The launch of these AI tools and training guides comes as the Times is embroiled in a legal dispute with OpenAI and Microsoft, with the New York Times claiming that ChatGPT conducted training on its content without permission. Many other publications are also introducing AI in their newsrooms at varying scales, from spelling and grammar tools to generating entire articles.