I do not want to repeat a story about Tai, an artificial intelligence (AI) chatbot that provoked violent reactions, so Microsoft closed it a day after its launch. For those who don’t know the story, click on two links in the first sentence.

There’s no need to repeat a story because the Microsoft idea behind Tai was stupid, and all the reactions I’ve experienced have missed the point where Microsoft became completely stupid. If I missed a reaction that hit the target, I apologize.

Intelligence in the brain?

The point is that all those who dream of AI confuse brain-like functions with intelligence. They do not understand that human intelligence is not concentrated in the brain or the mass of all human brains on Earth, even if they are miraculously connected.

The idea behind Tai looks clever for those who live only in a fantasy where all humanity is focused on the brain and its functions. Microsoft thought that if they updated semantic computer codes with crowdsourcing of human power, they would come to an actual human simulation, human intelligence.

Intelligence and human phenotype

I’m sorry, boys! Human intelligence rests throughout the human body. In other words, human intelligence emerges from the entire human phenotype. I would advise all techno freaks, including the Microsoft development teams, to read some books from evolutionary biology (early works by Richard Dawkins) and evolutionary philosophy (Daniel Dennett) to leave AI misleading paths. (And you can find much more – of course, not in the bestseller sections.)

Tai’s failure is a perfect example of why AI can never work when developed by techno freaks. They closed Tai because they learned racism and fascism literally through nightly interaction with the Twitter community. The developers were surprised by this “negative development”. On the other hand, some commentators even suggest that some “ethics” should be introduced to AI to avoid malfunctions.

Ethics and human phenotype

Dear all: Ethics results from the brain’s attachment to a body (body, in this respect, means the entire human phenotype and the extended phenotype)! Brains not attached to the body would go wild, like Tai, and at the same speed. What we can read on Twitter is already censored (filtered) by the body of the one who is writing and even more by the enormous complexity of the adaptations the reader’s body imposes on his brain. That’s why people (at least most of them) don’t become entirely extremist when they read social media, and therefore, no machine that simulates human brains can be intelligent.

Let me repeat:

NO MACHINE, WHATSOEVER POWERFUL,  CAN SIMULATE HUMAN INTELLIGENCE.

AI and evolution

The only way to AI construct real is to somehow (how?) repeat the entire human evolution. Every cell of every human body rests on the evolution of all life on Earth and beyond and on the pervasive history of every sensory input from birth. The smell of an egg on 15 November 1993 at 9.03 am combined with the smell of the same egg one second later (unlike the previous one) and so on is just as important to me at this moment as the sight of a dolphin ten years, seven months, 3 hours, 5 minutes and 45 seconds later. All these body sensations (smell, sight, touch, taste, sound) make us who we are. It is an ethic not to touch the fire that prevents us from being burned. It’s an ethic not to beat a friend because we once felt that he was hit by a strong fist that prevents us from killing – and so on. It is, in principle, impossible for Tai to make this experience, regardless of what techno implants she would get in the next zillion years.

Dear AI developers and techno freaks: If I could understand specific evolutionary lessons, you could, too. You have to step outside the how-to-do-it repetition jargon of we-can-do-everything USA mythology. Life is a “bit” more complex than the majority worldwide teaches.

July 2024, afterscript

I wrote the above post in 2019 before the Chat GPT and other AI tools became abundant. Did I change anything? Yes, I corrected some spelling mistakes and typos. The message is still valid as it was back then. And it will remain despite all foreseeable developments in AI tools. 

But what about the revolution that quantum computers bring? Not long ago, I had a lively discussion with a physicist and complexity theory expert on various topics that interest us both. He is convinced that quantum computing will bring a quantum leap forward for AI. (Quantum leap forward for quantum computers is something that came to my mind when I wrote this post. Nice parallelism.) If what I wrote above holds, then he was wrong. Quantum compute AI tools are going to revolutionize AI tools, for sure, but not to a degree that would come close to human intelligence. So, I hold my assessment of AI for now and for the future.