Last week we have seen some insensitive administrators praise AI to graduates during commencement speeches. These folks have not read the room, not seen the mood change that happened worldwide during the last couple of months. People start to get sick of it. They are also worried about it. I myself am very worried about it. Who would not, if your own profession is under attack? (in my case it is mathematics). Education, research, creative and administrative work, the arts, even consulting are all in danger of being replaced. They are already devalued. There were school administrators who think they have to be “hip” and use AI to read names even from graduates. Even if this stunt had worked technically, what is the message sent to graduates? It is insensitive: students have officially become a piece of merchandise that needs to shipped in the most effective way. It also started at Harvard. I have done first year advising for more than 20 years. This year, the program has been announced to be canceled. It is now done by AI (this is not official but I learned about such efforts to have AI advising instead of real people already last year by students of mine). Fortunately, Conan O’Brian’s commencement speech from yesterday did not focus on AI (there was a joke that AI is used to grade AI work and then some consolidation that AI will not replace you). Knock on wood. Was a good choice by the way to have him give the speech. (He himself has once called himself the “last human host” in an Oscar’s monologue). It was a joke but it is a serious joke. We might be the last generation that has been able to do creative work that is acknowledged as such. Any research or art now comes with the danger to be seen as AI generated. Even art from the past is in danger. One recently has labeled work of Monet as AI generated and immediately trashed it as “obviously AI generated”.
The problem with machines taking over is that we can almost not avoid it. It needs skill now to avoid AI. Search is now already mostly AI driven. We are given answers without being sent to a website that has the authority or source. It is taking over the internet like cancer. It is pushed in our faces everywhere, also in entertainment. Horror stories have emerged like of authors writing 200 novels a year, or AI writers winning writing prizes. It has become really hard to distinguish real from fake. We all see the slop entering all the entertainment, especially social media. Bots write now for bots. In math, we see mathematicians writing papers using AI already. Open research problems can be solved by illiterate kids. It is a devaluation of creative work. It is chilling to see that personal computing, having your own computing tools to become more and more expensive, because the data centers are eating up all that silicon. AI already starts to feed itself (data centers) and not giving any more to humans (personal computing). We are reduced to entities that have more and more difficulty to afford our own data processing. Resistance seems futile.
An other big problem is that we do not need intelligence to use artificial intelligence. A monkey can push a button uploading the question without even being able to read it. Once you know how to launch an agent, you do not even have to micromanage the machine any more. School administrators thinking that one has to teach kids how to use AI have missed the fact that the kids have already figured it out in 2022. It becomes increasingly clear that AI is a cancer. It also is almost certain that we have to live with this cancer from now on. It is a bad Nash equilibrium. It is like in an arms race. You can not in part of the word not develop it more as otherwise, the other part of the world will take over. This is the nature of Nash equilibria. Every player loses if it avoids going towards the equilibrium. Overall, the entire population might go down the drain. In the case of nuclear arms races, there was at least an obvious threat, the immediate destruction of the planet which has so far kept things in a balance. In climate, it might be natural logistic growth phenomena which could save us in the long term. But as for AI, there is no force at the moment that slows the growth. It might be people eventually (if there is no work, no purpose, no motivation no reason to live any more), but then it might be too late. The real danger is nihilism. School kids already start to say this: why do I need to learn if all knowledge is on my fingertips? And who will fight AI in the future, if we have lost the ability to think, acquire, process and store information? There are already signs for the reversal of the Flynn effect. I think it might be much worse as tests are also shifting.
So, what to do? It is hard to say. I myself try to treat is like cancer. I try to live with it, manage it, put it in its boundary, try not take over my entire body. Keep longer times in a day away from machines. Check things fed by the machine. At the moment this is still possible. It needs already more effort to keep AI away in your life than to lazily embrace it. Avoiding it becomes harder and harder. There are silver linings: the company Apple, which has not pushed AI into the operating system is suddenly again the hip one. People have seen the train-wreck that copilot that Microsoft pushed into their operating system, even in programs like Excel or Notepad. Having probabilistic elements in your work environment is a disaster. Having agents poking through your life is a nightmare. I can not believe that there are folks who accept that the machine takes screenshots every few moments to see what your are doing. I can not believe that there are scientists who accept data to be accessed and modified by AI. Experiments need to be replicable. This requires determinism. Lets hope that one learns from these disasters.
