Google is exploring an AI tool that can turn supplied information into news copy and has presented it to several major publishers, according to a report from The New York Times. The proposal raises a practical question for newsrooms: where can automation help journalists, and where must human reporting and fact-checking remain central?
A writing assistant under consideration
The tool is internally codenamed “Genesis.” Google reportedly sees it as a personal assistant for journalists, handling some tasks so they can spend time on others. The company has pitched it to The New York Times, The Washington Post and News Corp., which owns The Wall Street Journal.
Google has framed its work with publishers as exploratory. A spokesperson told TechCrunch that the company is in the earliest stages of looking at AI-enabled tools, especially in partnership with smaller publishers. The spokesperson described possible uses such as suggesting headlines or offering different writing styles.
That framing places the tool in an assistive role. Google says journalists would choose whether to use the technology, and that the tools are not meant to replace reporting, writing and fact-checking. The company compared the idea with assistive tools it makes available in Gmail and Google Docs.
Publishers weigh usefulness against risk
The New York Times report said some executives who heard the pitch found it “unsettling.” Their concern was that the tool seemed to overlook the work involved in producing accurate news. That reaction points to a broader issue: generating fluent copy is only one part of journalism, while checking claims and establishing what happened require careful work.
AI may offer practical support for particular parts of a newsroom workflow. Headline options or alternate ways to phrase material could help with drafting. But any generated text still needs evaluation by people who understand the reporting and can identify errors or unsupported claims.
That distinction matters because news copy can appear credible even when it is wrong. The source article warns that AI-generated stories without fact-checking or thorough editing could spread misinformation. The value of a writing aid therefore depends not only on what it can produce, but also on how journalists review and use its output.
Existing uses offer a limited comparison
Some news organizations already use AI for narrow categories of coverage. The Associated Press, for example, has long used AI to generate stories about corporate earnings. Those stories make up a small fraction of its articles overall; journalists write the rest.
This example shows that automation in news is not necessarily an all-or-nothing choice. A tool can be used for a defined task while much of an organization’s work continues to be produced by journalists. The source does not describe Genesis as a deployed product or specify how publishers might use it, so its role remains uncertain.
Newsrooms are also considering the wider question of responsible AI use. NPR and Insider have notified employees that they intend to explore how AI could be used in their newsrooms. Google’s pitch arrives amid that broader examination, but the available details do not establish what policies or safeguards those organizations will adopt.
Accuracy depends on review
A recent example in the source article illustrates why oversight is central. American media website CNET began quietly producing articles with generative AI, and the move backfired. CNET had to issue corrections on more than half of the AI-generated articles; some had factual errors, while others may have included plagiarized material.
Some CNET articles now carry an editor’s note explaining that an earlier version was assisted by an AI engine and that a staff writer substantially updated it. The example shows how a newsroom may need to correct and disclose AI-assisted work after publication, but it also highlights the cost of relying on generated copy without sufficient review.
For Genesis, the central test will be whether it can support journalists while leaving the essential work of reporting and verification intact. Google says its tools cannot replace that role. Publishers’ reactions and CNET’s experience make clear why accuracy, editing and human judgment are central to evaluating any such system.