We reported on this popular tech news outlet earlier this week CNET The AI quietly published entire articles written by the AI for months without immediately making its authorship clear to readers.
unlike the robo-reporting used by news agencies such as Associated Pressthese articles – 75 and counting – are substantial financial explanations, not just fill-in-the-blank updates, and are written with a more powerful AI similar to OpenAI’s GPT-3 (in keeping with the general spirit of privacy around). project, CNET did not specify which artificial intelligence it used to extract the articles).
Many observers, including those in the media industry, were not happy.
“This is just the beginning,” he tweeted The Washington Post reporter Nathan Grayson responded to the story, “and the aggregation and explanation performed by artificial intelligence will undoubtedly result in lower quality work and fewer jobs.”
“I think about things like this a lot because someone got fired from a copy editing job because some people think AI tools can do it for you.” wrote another writer. Kotaku writer Luke Plunkett it’s just called a program “terrible.”
The mentioned articles are published under “line”.CNET Money Staff” makes it clear that human writers are its primary authors. They’re not. Instead, the content is “generated using automation technology” and then reviewed by a human editor. Know that if you click on the line and read a small disclaimer .
It’s a pretty volatile way to make a statement, especially for a well-known brand. Other major news sites, e.g AP or LA Timesexplicitly tag the author as a bot or reserve the article with a clear declaration of AI authorship.
CNET‘s system is also more complex than a simple bot. It’s one thing to use AI to simply push out automatic scripted updates, but to use it to create all the explanations and then tell your audience it’s almost a new vulnerability makes even ex-employees nervous.
“As before CNET employee, it’s incredibly frustrating and frustrating, but not surprising.” he tweeted Kyle Hyatt, who now writes for him Jalopnik. “When you’ve fired all your talented and dedicated writers, what other choice do you have?
There is probably a certain sense of shame surrounding the human replacement project CNET, an employee there said they weren’t even aware of the AI articles until our interview. The company has not yet responded to their questions Futurism or made any other statement regarding other means.
Although clearly aware of the conversation. After our story, CNET has removed the “staff” from its AI story flags and now publishes it under “CNET Money.” The disclosure still refers to a small drop-down description, but now it says “created using an AI engine” instead of “automation technology”.
This slight rebranding doesn’t seem to make any more sense than before with regard to the use of artificial intelligence. “CNET Money Staff” was completely misleading, but abbreviated “CNET Money” is no longer illuminating.
And finally, such explanations do not remove the harsh reality that writing used to be someone else’s job.
“Writing articles like this kept me going in my 20s” he wrote Brenden Gallagher, screenwriter. “Shameful act CNET.”
Advocates of using artificial intelligence in news see it as a way to free overworked journalists and reporters from the mundane, busy work of writing. But if technology continues to improve, it’s hard to imagine media executives stopping there.
Moreover, the current version of GPT-3 “cannot cope with journalistic work”, a CNET “ChatGPT is amazing AI, but human jobs are safe (for now),” a journalist wrote in an article last month. Gizmodo.
At the end of the day, it probably depends on your definition of “journalist”. AI may not follow up on leads CNET quite yet, but he certainly tries copywriters from the office.
The result has been updated to clarify details on previous CNET reports that GPT-3 could replace journalists.
More on AI: Microsoft is working on a deal to add OpenAI’s GPT to MS Word