BRUSSELS — European journalists are increasingly using generative artificial intelligence tools in the newsroom despite concerns the technology could spur fake news and erode trust, a new survey shows.

Just more than half of the 286 journalists surveyed in Belgium and the Netherlands said they used generative AI tools such as OpenAI’s ChatGPT, researchers at the universities of Ghent and Amsterdam wrote in a study published Wednesday. Of those who do use it, 32 percent reported using it weekly, while 14 percent use it daily.

The researchers said those numbers were “in line” with other industry studies.

However, the vast majority of those surveyed said they’re pessimistic about the technology itself: 85 percent believe AI will worsen fake news and 83 percent think it erodes trust in journalism.

The most popular use is for automated translation (45 percent), transcribing interviews (35 percent) and proofreading (30 percent). Common reasons for the use of generative AI are saving time, working more efficiently or seeking inspiration.

The application of generative AI “primarily supports existing tasks rather than transforming or reinventing journalistic processes,” the study pointed out.

Generative AI tools became mainstream after ChatGPT rose to global fame at the end of 2022. The AI models underpinning those tools are regulated under the EU’s Artificial Intelligence Act.

Under the AI Act, developers of generative AI models face obligations such as disclosing how their models are trained, how they abide by copyright law and how they mitigate the most serious risks, such as any possible loss of control over their models.