201

monosyllabic as a computer. So the computer might be able to fool us well (pass the

Babbage test), but it won’t become a living being from that either, but only with evolution

and active representation of the environment (which after all might be possible to repro­

duce later).

14.2

Moreover, a computer can also be designed stochastically. For example, efforts to care

for the elderly in Japan are going towards making computers as human-like as possible in

appearance, responses, etc. This is called emulation, and it can emulate a human very well,

but with normal programming it has the same limitations as mentioned above. But you can

also equip computers with neuronal networks and let them gain experience in an artificial

environment and equip them with drives, etc. And if you add a certain amount of insecurity

to it, you can also use it to simulate a human being. And if a certain fuzziness is added, the

whole design is already very close to a living being. This means that the fundamental lim­

its for formal systems can be overcome more and more easily. However, the problem is

then exacerbated as to how a relationship can be formed between these artificial, increas­

ingly human-like machines and the people concerned. Again, a human (and not machine-

like) solution can only be found if humans remain self-aware and specify human values.

If the computer then reproduces itself, it is a real living being. The only thing missing

then is the ability to evolve – but that’s exactly what we don’t want, not only because of

security concerns, but also because of ethical concerns.

Current Examples of Artificial Intelligence

The most famous example of artificial intelligence and computer successes are probably

chess computers, especially Deep Thought’s 1988 victory over Grandmaster Bent Larsen,

losing to World Champion Garri Kasparov in 1989. Deep Junior won the 2011 and 2013

World Computer Chess Championships, and played Garri Kasparov to a draw in 2003.

14.2  Artificial Intelligence