2024 in AI weirdness: ‘Insane’ ChatGPT and mutant rat genitals were only the beginning

Before the waning year recedes into history, let us pause to appreciate that the past 12 months have brought numerous exciting advances involving AI and related emerging technologies. And in doing so, let us not fail to take a short last look at some of the more spectacular flops. 

It’s a fitting year-end exercise. Besides, some of the best of the worst moments delivered doses of that safest and most efficacious medicine: laughter. 

Ars Technica senior AI reporter Benj Edwards lays out a baker’s dozen of his favorite 2024 news items in a piece posted Dec. 26. In introducing his selections, Edwards suggests the “weirdness” unifying the examples may owe much to the novelty of new technologies.  

“Generative AI and applications built upon Transformer-based AI models are still so new that people are throwing everything at the wall to see what sticks,” he offers. “People have been struggling to grasp both the implications and potential applications of the new technology.”

Here’s a sampling of Edwards’s selections for 2024’s low-end AI highlights. 

1. ChatGPT went insane. 

Early in the year, things got off to “an exciting start when OpenAI’s ChatGPT experienced a significant technical malfunction that caused the AI model to generate increasingly incoherent responses, prompting users on Reddit to describe the system as ‘having a stroke’ or ‘going insane,’” Edwards writes. “During the glitch, ChatGPT’s responses would begin normally but then deteriorate into nonsensical text, sometimes mimicking Shakespearean language.”

From Edwards’s original coverage of the incident: 

‘It [was] like watching someone slowly lose their mind either from psychosis or dementia,’ wrote a Reddit user in response to a post about ChatGPT bugging out. ‘It’s the first time anything AI-related sincerely gave me the creeps.’

2. Mutant rat genitals exposed peer review flaws.

In February, Ars Technica senior health reporter Beth Mole covered a peer-reviewed paper published in Frontiers in Cell and Developmental Biology that “created an uproar in the scientific community” when researchers discovered it contained nonsensical AI-generated figures, Edwards reminds. 

From Mole’s February article: 

The figures contain gibberish text and, most strikingly, one includes an image of a rat with grotesquely large and bizarre genitals, as well as a text label of ‘dck.’

3. Robot dogs learned to hunt people with AI-guided armaments. 

“At some point in recent history—somewhere around 2022—someone took a look at robotic quadrupeds and thought it would be a great idea to attach guns [and other weaponry] to them,” Edwards recalls. “A few years later, the U.S. Marine Forces Special Operations Command (MARSOC) began evaluating armed robotic quadrupeds.” 

From Edwards’s April article: 

The Thermonator—what [private marketer] Throwflame bills as the first-ever flamethrower-wielding robot dog—is now available for purchase. The price? $9,420.

4. Google Search told people to eat rocks and glue cheese to pizza.

Google’s newly launched AI Overview feature “faced immediate criticism when users discovered that it frequently provided false and potentially dangerous information in its search result summaries,” Edwards writes. “Among its most alarming responses, the system advised humans could safely consume rocks, incorrectly citing scientific sources about the geological diet of marine organisms.”

From Ars Technica’s May coverage: 

Some of the funniest examples of Google’s AI Overview failings come, ironically enough, when the system doesn’t realize a source online was trying to be funny. An AI answer that suggested using non-toxic glue to stop cheese from sliding off pizza can be traced to a mischievous online troll.

5. San Francisco hosted a robotic car-horn symphony. 

In August, San Francisco residents got a “noisy taste of robo-dystopia when Waymo’s self-driving cars began creating an unexpected nightly disturbance in the South of Market district,” Edwards re-reports. “In a parking lot off Second Street, the cars congregated autonomously every night during rider lulls at 4 a.m. and began engaging in extended honking matches at each other while attempting to park.”

From the August article: 

The absurdity of the situation prompted tech author and journalist James Vincent to write on X: ‘Current tech trends are resistant to satire precisely because they satirize themselves. A car park of empty cars, honking at one another, nudging back and forth to drop off nobody, is a perfect image of tech serving its own prerogatives rather than humanity’s.’

Get the rest straight from the source.