Google is testing whether artificial intelligence (AI) can be used to write news articles – but insists it won’t replace journalists.
A spokesperson for the search giant said it was in the “earliest stages of exploring ideas” for how the technology could enhance “work and productivity” in a newsroom.
They cited coming up with headlines and suggesting different writing styles as examples of what AI could do.
“Quite simply, these tools are not intended to, and cannot, replace the essential role journalists have in reporting, creating and fact-checking their articles,” the spokesperson added.
But executives at publishers who have seen Google’s pitch said it was unsettling, The New York Times reported.
The paper was the first to reveal the company’s plans for AI-powered journalism on Wednesday.
It said it had held talks with Google, as had fellow US outlets The Washington Post and Wall Street Journal – plus Rupert Murdoch‘s News Corp, the publishing empire behind The Sun and The Times.
Asked about the report, a News Corp spokesperson said: “We have an excellent relationship with Google, and we appreciate [Google CEO] Sundar Pichai’s long-term commitment to journalism.”
How AI could change future of journalism
How Google uses news
Google relies on publishers for its Google News service, which aggregates articles.
Its search results are also often made up of stories from mainstream and local media outlets.
Last month, it announced it would remove links from Canadian publishers when the country’s government enacts legislation requiring tech firms to negotiate pay deals with them.
Canadian media has long complained that tech companies are piggybacking off their work and costing them advertising revenue. Google has argued the new law is a “link tax” that would “break” the internet.
Facebook owner Meta also opposes the law and is pulling links to Canadian outlets from its platforms.
Read more: Transparency ‘crucial’ over how AI is trained
Are AI newsrooms the future?
The New York Times report comes after the Associated Press, one of the world’s biggest news agencies, revealed it would partner with ChatGPT creator OpenAI to explore how it could use generative AI.
Earlier this month, Sky News also experimented with whether AI could perform journalistic tasks, including pitching and writing its own news stories, with mixed results.
Other outlets have provided warnings as to the risks, with technology news site CNET forced to admit it had used generative AI to write articles after embarrassing errors were pointed out in published stories.
The Irish Times also had to apologise for publishing an opinion article generated by AI.
Professor Charlie Beckett of the London School of Economics’ JournalismAI initiative, which helps outlets looking to leverage AI responsibly, has said “a lot of newsrooms are thinking through” what to do with the tech.
But “if we all get lazy and expect ChatGPT to write our stories and scripts, they may get worse”, he warned.