[ExI] ethics vs intelligence

[Anders Sandberg]

On 11/09/2012 21:13, spike wrote:
> Come on extro-ethics hipsters,
> think hard, suggest the right answers, and while you are at it, do again
> make the case that ethical behavior and intelligence are related please?
There is a link *in humans* between smarts and having a lower discount 
rate: you value the far future more if you can foresee consequences 
better. Smarter people also are better at cooperating in prisoner's 
dilemma situations since they can figure out the value of cooperation. 
But as we have been arguing from our horizon, this is no guarantee that 
super-smart systems will be nice.

http://www.nickbostrom.com/superintelligentwill.pdf
http://lesswrong.com/lw/cej/general_purpose_intelligence_arguing_the/
http://philosophicaldisquisitions.blogspot.co.uk/2012/04/bostrom-on-superintelligence-and.html

> Heh. All ethical dilemmas seem to pale in comparison to those 
> presented to the families of Alzheimer's patients. For instance, 
> imagine an AD patient who seems partially OK some mornings for the 
> most part, but nearly every afternoon and evening tends to grow more 
> and more agitated, confused, lost, terrified, angry, worried, 
> combative, clearly not enjoying life. But the patient sometimes has a 
> good day, and on those occasions clearly states a preference to stay 
> in their own home until there is nothing left of the brain. When is it 
> time to check the patient into elder care? Easy, right? OK what if the 
> patient's spouse is doing something wrong in the medication, such as 
> giving the patient large doses of useless vitamins, on pure faith 
> since Paul Harvey said they are good for this or that? What if you 
> come to suspect the patient is receiving sleep aids in the middle of 
> the day, and the rest of the family doesn't know? What is the right 
> thing to do there? Ignore one's own suspicion and go along, knowing 
> that if a patient is suffering, well hell, it isn't suffering to be 
> asleep, ja? Apparently AD doesn't hurt in the sense that it causes 
> pain, so it doesn't keep one awake as something like arthritis would, 
> but the suffering is real. If a spouse decided the person is better 
> off sleeping most of the time, is it appropriate to second-guess that 
> spouse? 

If I were to pretend to be a proper practical ethicist I would reason 
somewhat like this: There are some moral principles that are fairly 
robust, they emerge almost no matter what normal ethical system you use. 
In medicine the "classic" list is
<http://en.wikipedia.org/wiki/Medical_ethics#cite_note-7>

  * Respect for autonomy
    <http://en.wikipedia.org/wiki/Autonomy#Medicine> - the patient has
    the right to refuse or choose their treatment. (/Voluntas aegroti
    suprema lex/.)
  * Beneficence <http://en.wikipedia.org/wiki/Beneficence_%28ethics%29>
    - a practitioner should act in the best interest of the patient.
    (/Salus aegroti suprema lex/.)
  * Non-maleficence - "first, do no harm" (/primum non nocere
    <http://en.wikipedia.org/wiki/Primum_non_nocere>/).
  * Justice <http://en.wikipedia.org/wiki/Justice_%28ethics%29> -
    concerns the distribution of scarce health resources, and the
    decision of who gets what treatment (fairness and equality).

And then people often like to add respect for the person (sometimes 
solemnly dressed up as human dignity) and truthfulness as other key values.

Now, AD is bad for autonomy. Not all decisions are made in a rational 
way or using all available facts - but as you say, it varies from day to 
day. So clearly one needs to listen to what the patient wants. But 
sometimes the beneficience principle gets stronger weight. And when in 
doubt, one needs to be careful not to do harm.

Note that AD doesn't break up personal identity. Sure, memory is being 
lost and personality changes, but it is not as if the person is a 
sequence of independent identities. Hence earlier sensible decisions 
about one's future still hold moral weight and should be respected: if 
they want to stay at home as long as possible, fine. (That might of 
course be limited by the rights and considerations of the family: one 
should not follow all commands of a sick family member, no matter how 
loved. Other family members also need to have lives!)

The vitamin pills are likely not harmful (unless they are very 
expensive), so if that keeps the spouse happy that they are doing 
something or even provides placebo, there is little harm there. One 
might quibble about whether accepting one near-superstition opens up for 
other, more harmful remedies or bad epistemic practices, but it seems to 
be relatively minor.

The sleep aids, now that is different. That seems to break the autonomy 
principle rather badly, even if it is for the good of the patient. And 
if they are given secretly, then there are issues both with truthfulness 
and risk: some sleep aids might interact with other medications. It 
would be much better to discuss this with the patient when they are 
lucid and get them to explain what they want. This might be harder than 
it seems, because over time paranoia and other personality changes might 
make them want something different, at which point hard decisions about 
whether they are still capable enough to overrule past decisions come 
in. And talking it over with the spouse is of course hard in itself 
(still, one should not underestimate the burden of keeping things secret 
from one's own family).

In this case I think the principles do a bit of useful work, but 
obviously far more hinges on being a sensible person who can manage to 
talk about hard things in a sad situation.

Armchair ethicists might of course suggest more radical things (the 
Kantian: lying is *always* wrong! the utilitarian: maximize happiness, 
go for the sedatives and happy pills! the contractarian: follow whatever 
rules the AMA has published!) but for some reason they are not very 
common in real moral situations (no meta-ethicist in a foxhole?)

-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20120912/ebc955a0/attachment.html>