R2D2, robot character featuring in 'Star Wars' (Der Spiegel 2015)
What is intelligence? |
| Humans need company. Be it his / her best friend, the dog or the cat, be it other humans.
Also horses, parrots and crows may socialize with humans. But when it
comes to cognitively more ambitious feats as e.g. filling out an
online form or opening a bank account, we need more than just company.
In this regard, we highly appreciate to be treated by fair means,
without prejudice or bias, and to be guided surely across the turbulent seas of bureaucracy and malice. |
| Here intelligence
in its purest form jumps in. We call it 'artificial' (A.I.), because we
reproduce
in it our human talent to think and reason (sapiens) on the firm
logical ground of the sober naming and comparison of facts and the
neutral weighing of arguments, stripped bare of all subjective
interest. A.I. along these lines is a true friend of all
humans, respected and highly esteemed. When it comes to any formal
procedure, I prefer the truely intelligent robot to a
(fallible) human. It shouldn't even look human (I would welcome R2D2). |
| But
what about our company? Some of us don't like the idea of A.I. as a simple input / output machine
yielding correct answers. Maybe they
themselves do not really understand, why. They do not understand,
because their hope and expectation was a secret they didn't dare to
speak out frankly: they hoped for friendly company, for a
truthful partner, even with charms and beauty. But their secret hope
was without chance for fulfilment: there is no such thing as
artificial 'charms and beauty'. |
| Human company cannot be constructed. It has to be searched for, and it can be found.
It cannot be put together from boards and cards, with skrews and soldering. It
will not help to hide this fabric with enchanting textures
reminiscent of human bodies. Already in 1979 Gary Numan asked the crucial question: Are friends electric? They are not and can never be.
Friends have to be like us: with virtues and failures. We accept them
on our own level. Me and my friend, we both may be intelligent too.
Even in the dog I meet on the street I can sense it: the capability to follow its own thoughts. |
| The noun 'intelligence' may signify several abilities or predicaments. Already the Latin verb intellegere had at least two meanings. It meant the act of recognizing something, in terms of differentiating it from other entities.
A second meaning was related to understanding and comprehension,
including the particular meaning of a word or utterance. We may add as
a third the aptitude to draw conclusions, making use of two
different trains of thought by comparison. |
| The terms 'autonomy' and 'intelligence' belong to different categories. While 'intelligence' may be regarded as terminus technicus
reflecting the physical qualities of an instrument, 'autonomy' has a
more personal flavor. Given this distinction, it is no surprise
that we readily combine the term 'artificial' with 'intelligence' but
not so readily with 'autonomy'. Artificial autonomy? Vade retro, satanas! Maybe that is the reason, why some critics demonize A.I. agents: they fear their (suspected) autonomy. |
| We
are trained to fear malevolent humans. But machines? The more complex
they get, the longer it takes to learn their secure handling. In the
beginning of this process, we may fear our own mistakes (or those of
others). We usually regard machines as heteronomous and only fear that
autonomous agents may use them against us.They 'do' what they've
been instructed to 'do'. They have no motivation by themselves to 'do'
anything. Machines are no agents. |
| When I called A.I. a true friend of all humans, I did this in a metaphoric way ex negativo: A.I. has no own agenda to follow. It
is totally resistent to bribery. We can trust in A.I. It still has
knowledge deficits, but in short time it will know, what humans
globally know, and it will be able to communicate to us on all levels,
from the naïve child to the most sophisticated expert. And it will follow
rules, as we set them. And we are the good ones, aren't we? |
| We finally shall be. Thanks to A. I. /\ |
| 8/26 < MLB (8/26) > 8/26 |