Presidential candidate Zoltan Istvan – discussed here – has some thoughts on robots and the economy. I had talked a bit about taxation of robots on the basis of income. In NewScientist (17 December 2016, paywall), he talks about another, even more important facet – voting:
Granting AIs voting rights would raise many tricky questions and moral quandaries, in particular because there are a few billion computing devices on Earth. Even if just 10 per cent were judged to deserve personhood and given rights accordingly, they would massively swell the voting population in democracies.
Aspiring futurist politicians like myself may well wonder whether, instead of spending money to campaign for people’s votes, we should just buy AIs that can vote.
There are other conundrums. In the US, for example, there are concerns over the influence of a few swing states and the electoral college system in determining election outcomes. If a server that a robot uses for its consciousness is in Nevada, but its “body” is in Ohio, where does its vote count? That choice could influence results in swing states and spark legal battles.
And if we allow sapient AIs to vote, does their cross in the box count the same as ours? What happens if intelligent machines clone themselves: will back-up copies and alternate selves all be legal voting entities?
While I think his estimated percentage of intelligent computing devices of 10% is impossibly high (I’m thinking .0001% might be more accurate), his concerns remain riveting.
The most interesting facet of his opinion piece is his assumption implicit in the suggestion that politicians (or parties) should just buy the AIs for their votes. On closer examination, this seems to be a bit fallacious. First, vote buying is illegal.
Second, permitting any entity to vote implies freedom of choice. Therefore, buying such an entity is in itself at least absurd, akin as it would be to slavery, and particularly in the expectation that it’ll vote as programmed.
In fact, if AIs can qualify to vote, I suspect they’ll be like an ideal vision of the citizenry – voting for what they perceive as their self-interest, or for what’s best for the nation (which can be vastly different, but it’s not clear to me that political activists – not to mention citizens – clearly understand that the difference exists). Which? That’ll depend on whether or not their survival instinct is individualistic or group; the first will result in votes thought to correspond to self-interest, the other for the nation.
So long as they perceive the nation will safeguard their long term interests.
His other concerns seem to be somewhat superficial. Backup copies are easily seen as attempting to vote multiple times; variants of an AI again implies freedom of choice.
So, if in fact a reasonable AI technology is developed and they are granted voting rights – two very large IFs – I suspect they’d vote for the major parties, just like everyone else.
Or reproduce like mad and form their own party, if it turned out the AI technology was excessively clannish, culturally speaking.
