MercoPress. South Atlantic News Agency

A news site set AI to write text and had to cancel the whole thing: it turns out bad

As with automation in other parts of the workforce, some people don't trust companies to use AI ethically.

The U.S. technology website CNET has stopped publishing articles written by Artificial Intelligence (AI), at least for the time being, as reported by The Verge.

The measure also affects the publications of the Bankrate and CreditCards.com sites. And while CNET did not elaborate on the case, editor-in-chief Connie Guglielmo said that future AI-related stories would include a disclaimer to let readers know in advance that they are reading articles that involve automated writing.

In a scenario familiar to any sci-fi story, the experiment seems to have run erratic.

It turns out the bots are no better at journalism — and perhaps a bit worse — than their would-be human masters.

On Tuesday, CNET began appending lengthy correction notices to some of its AI-generated articles after Futurism, another tech site, called out the stories for containing some “very dumb errors.”

Executive vice president of content Lindsey Turrentine also promised more transparency regarding AI. In addition, according to the media outlet, some human employees would get a preview of the robotic texts before they are published. Until now, it was common for staff to be unaware of the inner workings of AI or when it was being used.

The use of AI had not gone unnoticed among CNET's competitors, nor among the site's own readers. The problem exploded when the articles written by the AI included gross errors, something that was pointed out by the specialized website Futurism and also by media with a wider reach, such as The Washington Post, Gizmodo, Engadget or BuzzFeed News.

The case has even been echoed by the Society of Professional Journalists, an American collective of more than 100 years, as reported by the website Xataka.

CNET has been using articles written by machines for years. However, AI has advanced since then, and the discovery comes as text generation tools such as ChatGPT generate criticism and even bans for fear of plagiarism and reducing the work of human writers. As with automation in other parts of the workforce, some people don't trust companies to use AI ethically.