Post B6IFmnc6tA898Zmh1c by LilaHexe@fedi.lilahexe.top
 (DIR) More posts by LilaHexe@fedi.lilahexe.top
 (DIR) Post #B6IFmnc6tA898Zmh1c by LilaHexe@fedi.lilahexe.top
       1 likes, 4 repeats
       
       My advice to incoming graduate students is very simple: AI is the Trillion Dollar Zombie, and extended use is likely to lead to your brain being eaten. The Zombie feeds in small bites, so it can be hard to detect, but every time you offload your decision-making and analytical powers to the machine, you are letting it take a nibble.You might think this is an extreme stance, far outside what we commonly hear in mass media. I think it is a reasonable way to treat a poorly understood public health threat, and AI is very much a threat to our well-being. While the technology is still young, we are already seeing in the news the awful effects on people in the tail end of the distribution of users. People who use AI heavily for counseling or romantic relationships tend to go right off the deep end. Their brains have been eaten. We even see effects on MathOverflow - people who are convinced that they have proved some big problem, and some AI has told them they are right and even helped to write a plausible-looking paper. Similar to other public health threats like heavy-metal poisoning, we initially only see direct effects with certainty on the tail end of the distribution, and normal effects are likely to be uncertain and depend on other susceptibility factors. This is not a reason to conclude that it is safe for normal levels of use. Indeed, we should learn from the history of how we have treated public health threats in the face of fat wallets, and listen more to people with small voices who sound alarms.(mathoverflow.net/a/511268)