AI in Perspective: Physical Limits of AI

For a good implementation of AI in society, all the social and ethical issues of AI in society need to be addressed. Otherwise even more issues will arise in the future.

The debate around the implementation of sentient supercomputers and complex algorithms into our daily lives often evolves into discussions regarding the ethics of AI. People have seen a massive increase in computing power and the applications of ‘smart’ devices in the last decades. A lot of people, scientists, and laymen alike are therefore certain that computers will soon outsmart humans in terms of self-awareness, empathy, and the ability to think critically. However, it is important to first think about where the limits to artificial intelligence lie and whether it is even possible to build consciousness into a computer.

Algorithms seem to be useful in almost every field nowadays, for example, machines have taken over a lot of operations the last few decades. Though the benefits for society of this change must not be underestimated, algorithms often require large data centers to function. These servers use a lot of electricity, so we need to carefully consider whether implementing AI is worth it. Perhaps we need to find new ways of producing energy for the computers they run on.

The question of whether it is possible to create artificial consciousness can probably not be answered without considering the biological and philosophical views on the subject. However, strictly speaking from a physics standpoint, there are two options. It is possible that all that is required to make conscious computers is a clever algorithm or series of protocols, which can run on devices that already exist or are close to being designed. We have a considerable chance of developing sentient AI if that is the case.

But a lot of people believe that we need far more computing power than we can create now to simulate a human brain, we might run into a problem. According to Moore’s law, the number of transistors we can fit in a circuit doubles every two years, but it appears that we are approaching the limit. This restriction is imposed by the size of atoms since we cannot build transistors smaller than the materials they are made of. We might not be able to create any sentient computers if no new way of building stronger computers is found. But these will undoubtedly have limits of their own.