Auch KI-Modelle bekommen Gehirnfäule | Eine neue Studie zeigt, dass die Fütterung von LLMs mit minderwertigen, hochinteressanten Inhalten aus sozialen Medien ihre kognitiven Fähigkeiten beeinträchtigt.

„AI models may be a bit like humans, after all.

A new study shows that large language models fed a diet of popular but low-quality social media content experience a kind of “brain rot” that may be familiar to anyone who has spent too long doomscrolling on X or TikTok.

„We live in an age where information grows faster than attention spans—and much of it is engineered to capture clicks, not convey truth or depth,” says Junyuan Hong … “We wondered: What happens when AIs are trained on the same stuff?”

Hong and his colleagues fed different kinds of text to two open source large language models in pretraining. They examined what happened when the models were fed a mix of highly “engaging,” or widely shared, social media posts and ones that contained sensational or hyped text like “wow,” “look,” or “today only.”

The models fed junk text experienced a kind of AI brain rot—with cognitive decline including reduced reasoning abilities and degraded memory. The models also became less ethically aligned and more psychopathic according to two measures.“

Can someone expand a little bit for me on what the term „cognitive ability“ means specifically when applied to an LLM? Is it meant mostly by way of analogy, as a term of convenience?

My understanding was that they are like a glorified autocorrect, going mostly by statistical probability. Is there a rudimentary reasoning of some kind as well? If so, could you characterize it with an example of the kind of rule that would be involved in simulating a reasoning process? I’m intensely curious about this.

Thanks!