CNET will stop publishing articles written entirely by robots after receiving a copious amount of negative attention over the practice during the last few weeks.
The affirmation was made on a conference call with editorial employees and executives at CNET’s parent company, marketing firm Red Ventures, on Friday, about two weeks after the website Futurism exposed several AI-written articles on financial topics that contained severe, glaring errors.
On Friday, CNET’s editor-in-chief Connie Guglielmo said the publication’s use of robots wasn’t done “in secret,” but was instead done “quietly,” and affirmed CNET disclosed their use of artificial intelligence to readers on the affected articles.
But that disclosure wasn’t initially visible to readers unless they clicked on an article’s byline. In most cases, the byline read “CNET Money Staff,” and there was no visible affirmation that the story being read was written by a robot.
Clicking on the CNET Money Staff byline resulted in a pop-up message that affirmed the story was “content generated using artificial intelligence.”
“A dedicated team of editors oversees the automated content production process, from ideation to publication,” the note said. “Ensuring that the information we publish and the recommendations we make are accurate, credible and helpful to you is a defining responsibility for what we do.”
But that process wasn’t enough to catch several egregious errors made in a number of stories that were ultimately published. One story misled readers on the issue of compound interest; another wrongly told readers they would pay a specific amount of interest if they took out a car loan with a particular rate. The errors were first reported by Futurism, the publication that exposed CNET’s use of content-creating robots.
On Friday, executives at Red Ventures affirmed CNET was not alone in using robots to create stories: Several other publications owned by the marketing firm, including CreditCards.com and Bankrate, also used artificial intelligence to write and publish stories. Tech website The Verge, which first reported on the call, earlier noted that some Red Ventures websites had used robots to write stories for more than a year.
CNET was one of several publications acquired by Red Ventures in late 2020 as part of a $500 million deal with CNET’s long-time parent company, Paramount Global. The agreement required Paramount Global, then operating as ViacomCBS, to lay off around 10 percent of its employees at CNET and other publication before the deal closed, The Desk reported. The layoffs amounted to around 100 cut positions across the acquired brands, which also included Gamespot, TV Guide and ZDNet.
Fewer editorial employees to oversee the robot-written stories at CNET and other ventures may have been a key reason why errors in those stories were visible for as long as they were until reporters at other publications began pointing them out. On Friday, some executives were critical of how Red Ventures’ use of technology was being covered by the media, saying reporters occasionally conflated the use of robots to write content with automation software that inserted stock prices and interest rates in some articles.
“Some writers, who I won’t call ‘reporters,’ have conflated these two things and had caused confusion and have somehow said that using a tool to insert numbers into interest rate or stock price stories is somehow part of some — I don’t know — devious enterprise,” Guglielmo said sarcastically on Friday’s call, noting that news outlets like the Wall Street Journal and Forbes have long used software to automatically insert stock prices into stories on certain companies that they are covering.
Lindsey Turrentine, CNET’s executive vice president of content, said the criticism over the website’s use of artificial intelligence will eventually blow over, but she declined to answer questions from employees who had concerns about erroneous data used by artificial intelligence software and possible plagiarism in some robot-written stories.
“This will pass,” Turrentine said. “We will get through it, and the news cycle will move on.”
Red Ventures has a very good reason for hoping the tech and media world will have a short attention span on this issue: Using robots to create content is central to the company’s overall mission of making money from readers through affiliate commissions.
Like other companies, Red Ventures uses affiliate programs connected to various products, including banking services, credit cards, streaming television services and even physical goods sold on Amazon. Red Ventures isn’t alone in this practice — other websites, including this one, use a mixture of advertising and affiliate links to generate revenue — but it has become one of the more-successful content companies to use it, building it into a $2 billion per year company, according to a profile written by the New York Times.
Some of the content written on CNET and other Red Ventures-owned websites — which used to include the websites TV Guide and Cord Cutters News — is explicitly created to appear higher in Google’s search results, which uses a proprietary algorithm to prioritize links to certain websites. Gaming that algorithm also isn’t new — there’s an entire industry built around so-called “search engine optimization” — and it’s one Google and similar directories have struggled with for years. But it can be incredibly lucrative when the first thing a person sees in a Google search is a link to an influential article on a Red Ventures-owned website that pushes readers toward certain products that earn the company money.
Done correctly, stories come across as just another news article or review column and the intent behind the content — to get readers to buy things — might not be obvious. The company stands to make more even money if it reduces headcount while increasing automation techniques to churn out more content focused on drawing readers in from Google and other platforms.
That appears to be the approach Red Ventures has used for a while now, and it’s the main point of frustration among journalists who have worried that the information industry will transition from reliable news toward retail operations.
It comes at a time when confidence in news is already at a staggering low among readers and viewers: Only 16 percent of Americans said they have a great amount of confidence in newspapers, and the number is even lower for viewers of television news, according to a Gallup Poll released last year. If the robots can’t even get things right, and if once-reliable platforms like CNET continue to be exposed as digital content sweatshops that place greater emphasis on quantity and sales over quality and reliability, that confidence will probably sink even lower to the point of non-repair.