Unless you have been living under a rock, you are aware that something big has happened in the world of Artificial Intelligence, just in the last month. I’m speaking of the release of ChatGPT4, which now can pass any bar exam, an engineering job interview, and complex math tests. It also can “get” jokes and has some insight into others’ “state of mind.” It passes the Turing Test hands down. As a result, at least one Paper has asserted that it shows “sparks” of General Artificial Intelligence, and people are making predictions that we’ll have GAI in as little as one year from now, and that this could lead to the “singularity,” where artificial intelligence uses its own intelligence to boost its intelligence, then uses its boosted intelligence to boost itself some more, ad infinitum until it is infinitely smarter than we are. Some say that we are already “in” the singularity, since AI researchers are using AI to improve AI, but I don’t think that’s quite right as a matter of terminology. For me, the singuarlity is when runaway improvement happens, and happens essentially in the blink of an eye. But what do I know?
And ChatGPT5 is set to come out next December.
Remember, ChatGPT is just one kind of artificial intelligence — it’s called generative AI because it uses large language models to generate text that comes out coherent and generally — although far from always — accurate. Another kind of AI is Alphafold, which basically can predict the folding of any protein that you throw at it — a feat that has taken humans several years per protein until now. This will obviously revolutionize biological research and developemt. Alphafolds forbears are also impressive — AlphaGo trounced the greatest human Go player, and AlphaZero trounced Stockfish, an AI that was previously the word’s greatest chess player.
And then there’s the AI that discovered Halosin, an antibiotic that works against a specific kind of bacteria. The scientists running it have no idea how AI knew that Halosin would work, but it did.
In other words, even apart from ChatGPT, AI outperforms humans at tasks that had once been thought to be uniquely human. It is more creative than humans in chess, and go and it sees connections in other areas that humans wimply can’t see.
And then there is AI-generated art, which can be quite beautiful.
So things are going to be changing real fast in the coming years, and it’s pretty clear that while some segments of humanity will likely do just fine, others are going to take a pretty big hit.
So there are at least five things we’re going to have to watch out for with AI. First, even seemingly benign uses will have built in biases that will result in unfairness. Second, corporations are going to use it in less-than-benign ways to make money off it. Third, bad actors are going to use it in positively malignant ways, which could spread large-scale chaos and destruction. Fourth, countries will use if for miltary purposes, likewise causing havoc. And fifth, there is definitely a non-zero chance that when AI becomes super-intelligent, it will enslave us or kill us.
Which is why we need a religion that focuses on our common humanity, is not particular to one racial group, and is based only on scientifically verifiable facts, where those facts are actually more miraculous than the greatest miracles in the rest of the world’s religions. Unless we develop and spread such a religion — whose ultimate goal is to venerate and preserve the human race — we may not survive the rise of the machines.
….to be continued