The film M3GAN was superb.

On the flight back from Vancouver to Tokyo after the TED conference, I watched the film M3GAN. Although I knew the hype about the film, I was not sure whether I had the guts to see it (I am weak about horrors and thrillers). However, after watching the film Nope (which is kind of a horror film, but I tend to accept sci fi), I felt that I was in a condition to watch M3GAN.

The film was superb. The script was clever and addressed some critical issues of AI alignment. M3GAN stands for Model 3 Generative Android, and the risks coming from uncontrollable generative artificial intelligence was well depicted in the film. 

This was days before ChatGPT surprised the world, and the producers of the film could be lauded for the future.

The mention of "attachment theory" in the script is symbolic of the general high standard of scientific and technological aptness. I recommend the film to anyone interested in the future of the alignment of AI with humans.

I am attending #TED2023 in Vancouver, Canada.

I am attending #TED2023 in Vancouver, Canada.

Despite competitions from likes of Lex Fridman podcast, TED still shines as one of the world's foremost market for exchanging ideas. Attending Session 1 I think I came to realize some reasons why.

It is the wholistic humane approach that characterizes TED. It is great to have geek talks about LLM and AGI, but at the end of the day, in order to alight AI with human society we need to value things outside the tech world. At TED we have a broad spectrum of people concerned with bringing the best in humanity. The better angels in us are flying in the arena.

It would be important to consider evolution of artificial intelligence systems in terms of group dynamics

For now, Artificial Intelligence systems seem to be developed as stand-alone entities, while historically,  evolution of biological species happened in the ecosystem. Instances such as ChatGPT are conceived as proving and being expected to prove excellence on its own. Dependence on the corpus makes them embedded in the ecosystem, though.

It would be important to consider evolution of artificial intelligence systems in terms of group dynamics, as happened in the case of biological systems, as opposed to the arms race metaphor typically employed in discussions about AI today, e.g. in the perceived confrontation between Elon Musk and Sam Altman, for example.

Related video. Towards a society of artificial intelligence.

Many things in Japan would perhaps go the way of Hachiko.

In Japan, for now, wherever you go, there are a lot of people. On the Shinkansen train, on the streets, at tourist attractions, everyone everywhere all at once.

With the covid-19 pandemic, the tourism came to a standstill, as in many countries. Now tourists are back with a vengeance. A noticeable change is the presence of people from abroad compared to domestic tourists.

Perhaps this is a vision of the future for Japan. In comparison to other global destinations such as London, New York, and Paris, Tokyo feels like and is still a place where the effects of globalism is seen only in mild signs, unless, of course, you go to the Shibuya Crossing. The Hachiko statue nearby is now always busy with tourists from abroad queueing up to have a chance to take a photo beside the famous Akita dog.

When I was a college student I could not imagine a day when the world would come to meet and greet Hachiko. It was a domestic presence then. Now many things in Japan would perhaps go the way of Hachiko. What a time to live in, apart from the rapid development of artificial intelligence. 

High intelligence is a double-edged sword

The conventional wisdom would be that if you have high intellect you would be more adaptive to a wide range of environments. Homo Sapiens has evolved to possess a highly developed intelligence, and it surely correlates with the fact that humans have come to dominate in a wide range of environments, from the tropics to the north and south poles on the earth, and to International Space Station and further beyond, perhaps even to Mars.

However, although intelligence has surely helped humans to be more robustly adaptive in a wide range of environments, it has also made the human existence less robust and stable. The possibility of human extinction through total nuclear war is just one example.

It could be argued therefore that high intelligence is a double-edged sword. On the one hand it can help make the system more robust. On the other there would be increased vulnerabilities, easily scaling out of the comfort zone.

It is an interesting question whether incorporating artificial emotion or consciousness in a system would make it more or less robust. Memorably, Eliezer Yudkowsky remarked in a recent Lex Fridman podcast that endowing an AI with emotion would be terrible. Artificial consciousness might make a AI more stable, by the incorporation of metacognitive processes, realizing the veto function, which is indispensable in human ethics.

The jury is still out.

In a recent episode of Ken Mogi's Street Brain Radio I discussed these pressing issues in some detail.

Related video:

High intelligence, artificial or natural, becomes unstable. Can consciousness help that? 

The arms race happen between people, not AI systems.

For some time now people have been discussing existential risks for humanity with the development of artificial intelligence. Although there would be genuine vulnerabilities due to the general disruption that the intelligence-related technologies would cause, especially by those involved in military operations, the tendency to depict AI, AGI in particular, in the light of possible overtaking of human existence is not only misleading but also potentially damaging.

Typically, when people discuss doomsday scenarios, they are projecting their own psychology onto the machine. It is not AI that would try to overtake the world. People have desires and ambitions about exerting control over others, and artificial intelligence systems are regarded as tools to realize their obsessions.

The arms race happen between people, not AI systems. The alpha male projection of aggression on the coming AGI is not only misplaced but also damaging to the neutrality of the technology.  

Related video.

The existential risk of Artificial Intelligence only comes from human nature and imagination

Saving Japan

In the last few years I have written two books on Japan. One on ikigai and another one on nagomi. With these attempts, I have hopefully presented the best in the tradition of the land of the rising sun.

As I have written in the small print sections of these books, I had no intention of claiming that Japan is the best, or indeed, unique among nations on the globe. Each culture has its own merits and strengths, juxtaposed with shortcomings and weaknesses. Japan is far from perfect, especially when it comes to gender equality, for example.

In a way, with the ikigai and nagomi books I have presented a vision of what Japan could be, could have been, and would be, in addition to what it actually is. I believe realities can be seen from a new and hope-giving perspective, when you have the perception and good will to achieve that.

I really admired the film Saving Mr. Banks. It told the true story behind Mary Poppins. As a lover of the excellent musical film, I believe in the alchemy of transformation from the actual Mr. Banks to the fictional character, depicting what he could have been, inspiring people. 

In the same spirit, I wanted to do something in the spirit of Saving Japan, while remaining true to the essential nature of the nation. Sometimes, you see the real self better from a distance.