You may have seen the story this week about an "AI" (LLM) bot that opened a pull request against the Python #matplotlib library. When the request was closed, the story was that the #bot posted aggressive condemnations of the maintainer(s) in the thread, particularly Scott Shambaugh, and then posted an article to a blog attacking him.
There's a lot of speculation about how much of that action was driven algorithmically vs manually instructed by the bot operator.
Anyway, so far so interesting. Lots of online media coverage of it, because it was novel.
But then, Ars Technica published an article about it. And Shambaugh read it, and posted that the second half of the article contained alleged quotes from him that he never said. The article appears to have been "AI"-generated, without any acknowledgement of that fact, which is not good for Ars' credibility.
Worse, Ars then quickly pulled the #article - silently. No mention of it left, no #apology to Shambaugh, no note to the readers of what they'd done.
That is absolutely #unethical in #journalism. It is unforgivable.
Ars was once a technical site with additional coverage of science and other geek-adjacent topics. Then it went commercial, was bought by a huge publishing conglomerate, and became a lot less technical.
But this is a slide too far. Unless they fix this, quick, I won't renew my subscription.
Mirror of yanked article:
https://mttaggart.neocities.org/ars-whoopsie
#ArsTechnica #ars #ethics #PR #LLM #AI