The Singularity of Stupidity
I hope you have a parachute, because the cliff edge is approaching fast
“I found it on the internet, so it must be true!”. This meme is becoming a bigger joke than ever, with the latest estimates that 52% of articles on the internet are now AI-generated. Personally, I suspect it is higher than that if LinkedIn is any indication, and will certainly climb as more people become dependent on AI to read, write, and research information.
So what’s the issue with that? The problem is that AI is not generating information a priori; it is doing so based on information on the public web, which is already AI-generated. Rather than increasing the base of knowledge, we are instead curating and concentrating it as AI generates more material based on AI-generated material, which in turn is used by AI to refine their models, which in turn generates more information… You get the picture.
Add to this the issue that LLMs do not know what is “true”. For that matter, they do not “know” anything. They are statistical sampling engines that create inferences from large data sets. When those data sets are themselves created by AI, the problem is that the pool of true knowledge rapidly shrinks as a percentage of the information available. Instead of knowledge, we are mass generating garbage.
Compounding this is the issue that the models themselves are not only prone to knowledge concentration, they are also prone to corruption by those who create and tune them, by the selective use of the data that they feed them as the sampling base. This has been more than aptly demonstrated by the leftist/woke bias in the responses to queries as documented for all the major AI players.
Finally, there is the stupefaction of humanity itself by its increasing dependence on AI to do the work that until very recently required that you rub your brain cells together to do. From grade school and college students to law clerks and politicians, we are rapidly losing our ability to think and reason, instead increasingly relying on the crutch of AI to give us the answers we are seeking. Who cares if the answer is wrong or biased? Look how quickly we got it!
I certainly have complained in recent years about the inability of Gen Z and Alpha to count change or do math without a calculator. God forbid that they actually have read a book in their lives. Now with them increasingly dependent on AI for not only answers and information, but also digital companions and relationships, it seems that we are speeding toward turning the movie Idiocracy into reality in our lifetime. Certainly, the increasingly public displays of emotion vs reason as a means of debate are just one indication of this trend.
So what’s to be done? I’m afraid I despair about anything that can reverse this trend. Like lemmings rushing to the cliff, humanity as a herd is trapped in group think and the inability to sense the danger we are in until we have already gone over the edge.
Can this be reversed? It certainly could, by strict laws and regulation on the uses of AI, as well as a wholesale reform of education by returning to the basics (reading, writing and arithmetic) as well as restrictions on the use of social media and computers by youth. Will it be done? I’m afraid there is not a snowball’s chance in hell, that being the destination we are rapidly approaching…
BTW, this article was not written by AI - yes, I actually do still have a functioning brain