English | EN 
Bennie Mols is a Dutch science journalist and author of the book Turings Tango, a book about the search for A.I. and the relationship between humans and computers. He graduated in physics (M.Sc.) and philosophy (M.A.) and holds a PhD in physics. He writes and thinks about the intersection of physical sciences and mathematics, computer science, neuroscience, and technology.
In his talk Mols takes the audience on a journey to separate facts from fiction in the field of A.I. First up is fiction and therefore singularity. The Singularity movement is based on Moore’s law and in the eyes of Mols a myth. Singularity is a myth because the amount of transistors on a chip might grow exponentially, but this is not the case for software, the clock speed of processors and the import and export of data.
The second argument is that Moore’s law does not solve the lack of intelligence. A cockroach is smarter than the best robot in the US army. We people are way better in learning and image recognition. The amount of calculations per second is not the same as intelligence, according to Mols. That’s also the reason that we not become obsolete because of robots.
That’s also a myth according to Mols. For two reasons. One reason is the paradox of automatization. The more machines take over intelligent tasks, the role of humans becomes more crucial in fixing their mistakes, adjusting performance, maintenance and improvement.
The second reason: robots will never be related to apes, like humans. Of course Mols talks about biology here. We humans have emotions, intuition creativity, values and rationality. Things that machines don’t have and the things that give us an biological advantage over machines.
Also, there’s this minor detail: a human brain can be compared to a energy level of 20 watt, IBM’s Watson (the smartest machines today) needs 6 MegaWatt, the equivalent of 300.000 human brains or 16.000 average households.
So in fact computers and humans have very different qualities:
- Computer: fast calculations with no error, perfect memory, super fast search, tireless and really good at exact science.
- Humans: learning, pattern recognition, social-emotion intelligence, good at handeling anomalies, multi-functional and creative.
- Mobile devices will use the big data and computing power of the web
- Computer will get better at perceiving reality
- Better knowledges of the human brain will give A.I. a boost
- There will be symbiotic networks based on people, computers, robots and things.
Thank god, we do not become obsolete!
But what about Watson, won in a direct game from the human and is better in diagnosis than the doctor …
Very thought provoking!
So one argument is: there isn’t enough capacity (/power/bandwidth/..) , and I think we’re seeing that capacity is never an issue that stops progress. We’re now putting massive compute power into everything we own, so collectively the world will have so much spare capacity that we could surely run a couple of ‘brain simulations’ at one point. We have plenty of energy, if we want to, etc.
The other arguments: emotions, intuition creativity, values, rationality (whatever that means) are pretty weak too, I believe – there is no proof just yet that emotions cannot be modeled, that ‘intuition’ is not in some way the outcome of a neural network that can be simulated or that creativity is beyond the reach of calculation. The fact that we don’t understand it just yet doesn’t mean we can’t understand it ever. And if we do start to understand the brain better, and for example learn which triggers cause a certain reaction in the human brain, couldn’t we start to simulate/copy this in a computer? I know human beings like to believe that our brain’s processes are beyond the realm of rational modeling, but why would it? If we can already stimulate the brain in certain areas to evoke a certain feeling, how can we say this ‘feeling’ is anything but bio-chemical in nature? With psychology and neurology we are becoming better and better at understanding what drives our emotions. And with this advanced understanding comes the option of simulating/automating..
I would not underestimate the exponential capabilities of technology…
I do like the idea of symbiotic networks, though somewhere in the back of my mind comes a little voice that says: “We are borg, resistance is futile”… 🙂