Months after the news broke that tech outlet CNET had started quietly by producing articles with generative AI systems, the site clarifies how it will and will not use the tools in the future.

To be below promises: stories will not be written entirely using an AI tool, and hands-on product reviews and testing will be done by humans. CNET will also not publish images and videos generated using AI from now on. But the outlet says it will “explore the use of AI tools” to sort and analyze data and create outlines for stories, analyze existing text and generate explanatory content. The tools for the home CNET is called Responsible AI Machine Partner or RAMP, according to the memo.

CNET has also gone back and updated the dozens of previously published stories generated using AI systems that caused a backlash in January. Of the more than 70 stories published over the course of several months, CNET eventually issued corrections for more than half. Some contained factual errors, while others were updated to replace “phrases that were not wholly original”, suggesting that they may contain plagiarized material. Sstories now add an editor with the text: “An earlier version of this article was supported by an AI engine. This version has been substantially updated by a staff writer.

Red Ventures, a private equity backed marketing firm that bought CNET in 2020 from CBS, has implemented AI systems in the many brands and websites it owns — Bankrate, a personal finance website, has also dozens published of AI-generated stories, just like Creditcards.com, another site owned by Red Ventures. Across all outlets, the Red Ventures playbook is the same: publish a bunch of SEO bait articles with keywords people are searching for, then load the pages with lucrative affiliate marketing ads. The company then profits when readers click through, open a bank account, or sign up for a credit card.

The AI ​​policy update comes just weeks after that CNETeditors announced they had formed a union with the Writers Guild of America, East — and guardrails around the use of AI systems were a concern. Employees cited a “lack of management transparency and accountability” regarding the use of AI tools, as well as concerns about editorial independence at the outlet. The policy was drawn up internally and the union was not involved in discussions. (Disclosure: The Verge editorial team is also affiliated with the Writers Guild of America, East.)

In February, The edge reported that CNET journalists were sometimes pressured to change their jobs to appease advertisers, and some staff members were asked to work on advertisements for Red Ventures clients, causing frustration and prompting staff to push back.

In a tweet, the CNET Media Workers Union said it would negotiate key issues such as testing and re-evaluating the tool and the ability to retrieve bylines before deploying the tool.

CNET is just one of many high-profile news outlets that have announced the use of generative AI models for storytelling. In February, BuzzFeed started publishing quizzes that used AI software to generate answers, eventually publishing dozens of travel guides using AI tools. In April, Insider said it experimented with using ChatGPT to generate SEO headlines, prepare and outline interviews, and add AI-generated text to stories.