[alicebot-aiethics] translation of Hawking, oh my god what is the fuss about?
Noel Bush
alicebot-aiethics@list.alicebot.org
Tue, 11 Sep 2001 18:09:24 +0400
> > Really? any mention of altering human dna is eugenics, and
> > therefore Nazi like? did I understand you?
>
> Yes. This is exactly what I mean. So far every single time
> we've (we=all
> humanity) tried formal eugenics programs we've only ended up
> causing great suffering and degrading the whole species in
> the process.
>
> Eugenics is a system where one path appears to lead to Utopia
> but where the other 100,000 paths lead to disaster scenarios
> beyond our imagination, such as the Holocaust or the Khmer
> Rouge. It's easy to think of dozens of disaster scenarios
> involving the rapid adoption of eugenic principles. Can you
> imagine what would happen if, for example, we had some
> yardstick for measuring one human's genetic value over
> another's? Einstein's gene scan would have revealed a little
> dyslexic kid with no respect for authority. Charlie Parker's
> would have revealed severe manic depression.
I agree with that. Anybody who thinks that the most important problem
we need to solve at the moment is how to make people live longer with
fewer diseases has mixed up priorities, I think. With respect to
disease, the bigger problem that we face is attitude to disease. Many
diseases don't have cures. People with those diseases face
discrimination. Putting eradication of disease at the *top* of the list
naturally leads to eradication of people with disease: after all, in
many cases that would be the most effective way to halt infection. Look
what we did with all those cows in the UK and elsewhere as a result of
mad cow disease.
And even discrimination against people with disease isn't the topmost
issue, if we're in the "planning for a better world" mode. A black
single mother living in poverty in some US inner city still has the odds
stacked against her when it comes to getting medical treatment, a job,
or just respect in a lot of places. A person of Turkish descent living
in Germany has it a lot tougher than a "native" German. And so on.
There are a lot of mythical beliefs that support people's tendency to
rank the importance of others based on the categories into which those
others fit. Some of the mythical beliefs might bear on AI.
When people describe what they expect to see in an artificial
intelligence, they are often describing what they would expect to see in
a "perfect" person. It is rare to find someone describing possibly
"acceptable" AI as being "flawed" in some way. Try to understand what
is implied to be a flaw in someone's vision of "successful" AI, and you
may uncover some pernicious biases.
One bias has to do with the way people speak. Somebody on the
alicebot-general list recently cracked a joke that it would be funny if
a bot spoke "ebonics". Ebonics as I understand it is an attempt to
recognize the linguistic "validity" of English as spoken by many black
citizens of the US. Naming this as a separate language or linguistic
category strikes many white Americans as silly, because most white
Americans (most people all over the world, for that matter) aren't aware
of the tremendous diversity in present-day English as spoken all over
the world. I'm sure you can find plenty of people who would visit
Jamaica and conclude that people speak English there "badly".
The "fuss" about improving education, improving healthcare, improving
general quality of life is often misguided in its assumptions about what
really constitutes "improvement". It's also often whitewashed into
terms of "socioeconomic classes", which conveniently skips over the
cultural, religious and ethnic barriers that really divide people, and
implies that if we could all be rich we'd all be okay.
If there is any interesting ethical angle to artificial intelligence, it
is in looking at the assumptions that people are working with when they
are trying to create AI. If there would be any value in an ethical
pursuit of AI, it might be in holding up a mirror to humanity and saying
"this is what we are". If that mirror shows an Ubermensch with
"unstoppable powers", an Ubermensch that is "above" race, religion,
bodily limitations, and so on, then it's every bit as destructive as the
image of an Ubermensch as a blond-haired, blue-eyed Olympic superstar.
If there is a fight to be had, it's against the people who set the
eradication of all "imperfection" as the highest priority. "Getting
along", and how to do so, is a much higher priority. Accepting and
accomodating all manner of physical or mental disability, cultural
difference, religious difference, and so on, is a far more challenging
and worthy goal. It is a goal that is compatible with "progress".