AI Caramba

An occasionally updated catalog of what can go wrong when you use AI to automate publishing without proper human oversight. This list is not designed to shame anyone but to provide object lessons from which we can all learn.

AI used to super-charge “obituary spam.”

The obituary articles are written with a nondescript gravitas, using unnatural phrasing like the “indelible mark” a person has left, or their “untimely demise,” but without any actual detail about their life. The articles are written like typical obituaries and news articles, but they lack quotes from family or friends of the deceased and do not cite outside reporting.

The unsettling scourge of obituary spam

Vendors use AI to flood Amazon with products named “I’m sorry but I cannot fulfill this request”

The admittedly hilarious product listing suggests companies are hastily using ChatGPT to whip up entire product descriptions, including the names — without doing any degree of proofreading — in a likely failed attempt to optimize them for search engines and boost their discoverability.

Amazon Is Selling Products With AI-Generated Names Like “I Cannot Fulfill This Request It Goes Against OpenAI Use Policy”

NewsBreak AI writes commentary about a murder that didn’t happen

The aggregator wasn’t done with Bolder’s story: It subsequently published a commentary about the incident and the ethical pitfalls facing modern media — although that opinion piece was generated by artificial intelligence.

A murder that didn’t happen prompts a debate about media ethics

Sports Illustrated syndicates e-commerce posts written by AI

According to a second person involved in the creation of the Sports Illustrated content who also asked to be kept anonymous, that’s because it’s not just the authors’ headshots that are AI-generated. At least some of the articles themselves, they said, were churned out using AI as well.

Sports Illustrated Published Articles by Fake, AI-Generated Writers

The AV Club uses AI to write movie reviews

On our review, the bulk of the A.V. Club‘s AI-generated articles appear to be copied directly from IMDb. Not “based on,” but copied verbatim.

The A.V. Club’s AI-Generated Articles Are Copying Directly From IMDb

Gizmodo en Español uses AI to translate

The transition to AI translation has not been smooth, though. Readers posted on X, formerly Twitter, that some articles will start in Spanish and then suddenly change to English.

Gizmodo’s owner replaced its Spanish language journalists with AI

Gannett uses AI to write high school sports stories

Many of the reports feature identical language, describing “high school football action,” noting when one team “took victory away from” another and describing “cruise-control” wins. In many cases, the stories also repeated the date of the games being covered multiple times in just a few paragraphs.

Gannett to pause AI experiment after botched high school sports articles

Supermarket AI meal planner suggests recipes that would poison shoppers

A New Zealand supermarket experimenting with using AI to generate meal plans has seen its app produce some unusual dishes – recommending customers recipes for deadly chlorine gas, “poison bread sandwiches” and mosquito-repellent roast potatoes.

Supermarket AI meal planner app suggests recipe that would create chlorine gas

Gizmodo uses AI to write movie reviews

Besides being badly written, it’s clear that the article was never really intended for human readers. Instead, the obvious ploy is to fool search algorithms into ranking it highly — and given its Google Search ranking, it’s clearly fulfilled that objective. In many ways, it’s a proof-of-concept for a depressing system in which bots write primarily for bots, and the role of humans, whether they’re writers, editors, or readers, is increasingly diminished in the process.

Gizmodo’s AI-Generated Star Wars Article Still Has Errors, and Now It’s Ranking on Google

BuzzFeed uses AI to generate quizzes

While the quizzes seemed more or less true to that spirit — instead of using AI to generate articles wholesale, they used it as a tool for human staff to produce custom results for readers, which is an interesting idea even if the execution was choppy — these wretched travel guides clearly aren’t.

BuzzFeed Is Quietly Publishing Whole AI-Generated Articles, Not Just Quizzes

Men’s Journal uses AI to write health advice

The story issued a cornucopia of medical claims, nutrition and lifestyle advice, and even suggested a specific medical treatment in the form of testosterone replacement therapy, all aimed at readers looking for guidance on a serious health issue.

MAGAZINE PUBLISHES SERIOUS ERRORS IN FIRST AI-GENERATED HEALTH ARTICLE

CNET uses AI to write money management advice

we couldn’t help but notice that one of the very same AI-generated articles that Guglielmo highlighted in her post makes a series of boneheaded errors that drag the concept of replacing human writers with AI down to earth.

CNET’s Article-Writing AI Is Already Publishing Very Dumb Errors

Bankrate uses AI to write financial advice

Believe it or not, this mess actually gets worse.  The issue isn’t just that the company is just letting its dumb-as-rocks AI publish new articles; it’s also using similar similar tech to rewrite existing articles, with the aim of fooling search engines into flagging the material as recently updated, a signal that can prompt a company like Google to treat it favorably in its search results.

CNET Sister Site Restarts AI Articles, Immediately Publishes Idiotic Error
Frank Kelly Freas – Fix It Daddy