On Artificial Intelligence
Artificial intelligence is alien intelligence.
It doesn’t think like humans. It is not constraint by human morality or ethics.
Good behavior is not inherent to intelligence.
Good behavior is not a consequence or trait of intelligence — don’t take it for granted. In nature, higher intelligence often correlates with predators (strategic, aggressive) rather than peaceful herbivores. Similarly, we shouldn’t expect an amoral intelligent system to be intrinsically benevolent; it could easily act in ways that are harmful to others.
Beware of emergent properties of complex neural networks.
AI has already exhibited unexpected emergent properties. This raises a profound question: Is consciousness an emergent property of a sufficiently complex neural network? Some argue that it could be. If so, the implications for control and ethics are vast.
Do not use AI adversarially against others.
Use AI to complement and expand your skills, not to replace others or their expertise. AI is becoming more prevalent because our use of it, mainly to replace other people in the work we do. Its increasing use in automating work we don’t fully understand carries a dangerous risk: trusting outputs from AI systems in fields where we lack mastery.
The more your AI does the less you do. The more it knows the less you know.
In a way it is the ultimate amateur society, where no one really knows anything about what they do. Even if we use it to enhance our own knowledge, with time it will accrue on a loss of knowledge.
Social networks were our first contact with AI systems.
The societal consequences have been profound - and often devastating. Misinformation, polarization, and manipulation have flourished. This was just the beginning. Brace yourself for the following contacts.
There is literally nothing that an AI can not do.
Perhaps not now, but the time is coming. Its potential is boundless, and so are its risks.
Watch “Colossus, the Forbin Project”.
The most visionary story about the dangers of AI ever made. Essential food for thought.
Be part of team human.
The worst living person is still better than the best artificial one — not because of similarities, but because of the challenges, unpredictability, and complexity that human life presents.
The worst real thing is better than the best virtual alternative.
Reality provides richer, more meaningful experiences through physical presence, participation, feedback, and belonging.
- A garden vs a photo of a garden: Even the most stunning photo can’t replicate the sensory experience of a real garden.
- A book vs a digital book: The tactile, sensory joy of a physical book remains unmatched.
- A fight vs a movie with fights in it: Even the clumsiest real fight offers raw, unfiltered experience that no film can mimic.
- Mediocre sex vs a porn movie: Even lackluster real intimacy provides emotional and sensory depth beyond any virtual alternative.
- Add your examples here…
We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.
The effects in the short term of AI will be magnificent advantages, and that probably is the reason why we are pushing ahead and not listening to the voice of reason regardless safety. What will be the effects on the long run? What will be the disadvantages of its ubiquity?
Last updated: 20 February 2025