Lesson for the 21st Century

August 31, 2018 6:36 AM

Why Technology Favors Tyranny - "Artificial intelligence could erase many practical advantages of democracy, and erode the ideals of liberty and equality. It will further concentrate power among a small elite if we don't take steps to stop it." (via)
posted by kliuless (8 comments total)

In the second decade of the 21st century, liberalism has begun to lose credibility. Questions about the ability of liberal democracy to provide for the middle class have grown louder;

No, sir, it isn't liberal democracy, it is unrestrained capitalism.
posted by grumpybear69 at 6:40 AM on August 31, 2018 [12 favorites]

Liberal democracy has not, historically, been exactly a bulwark against unrestrained capitalism. As the article says,
"Liberalism reconciled the proletariat with the bourgeoisie" after a fashion. Still so many of the suggestions I see for "socialist" reform hinge upon class collaboration and rarely centre the sole advantage we have as workers - the ability to withdraw our labour.
At least this article, even if it is jumping the gun as so many AI-related pieces do, recognises that if we lose that, we lose everything. If we're made economically irrelevant though, I don't know that the concern will be controlling billions. I don't think the ruling class will keep billions around if we are of no use as workers.
posted by AnhydrousLove at 7:21 AM on August 31, 2018 [6 favorites]

This piece's single-minded worship of efficiency is emblematic of the core problem. Much of the issue is that efficiency is pushed above all else, which is where harm comes in.
posted by NoxAeternum at 7:29 AM on August 31, 2018 [2 favorites]

I don’t know if “intelligence” is really quite the right term for a set of advertising revenue maximization algorithms that accidentally destroy democracy.
posted by Artw at 7:36 AM on August 31, 2018 [10 favorites]

Remember to separate the American colloquial term "liberal" meaning "anything to the left of the right wing" with the political science term of art "liberal" meaning "favouring laissez-faire economics".

This is a reminder to the author of this article as much as it is to anybody in this thread.
posted by tobascodagama at 8:07 AM on August 31, 2018 [7 favorites]

The AK-47 is sometimes held up as an example of a weapon that benefits the resistance as much as the oppressor, because it's cheap and robust and it amplifies the deadliness of each soldier. This compared to a cruise missile, which is the opposite: very expensive, requiring tremendous support infrastructure, and doesn't scale with the number of boots on the ground.

The open question is whether AI is an AK-47 or a cruise missile. As a computer scientist, but not an AI expert, I could make a good case for "it is an AK-47", because I know the power of things I could hack together with a raspberry pi and open source machine learning libraries in relatively little time, and that feels like something accessible to the resistance. But it would be disingenuous, because deep down I am quite certain it is a cruise missile, and that AI will amplify the infrastructure of oppression more than it amplifies the resistance.

My grand conclusion from the political events of 2016 onward is that "how to put good people in power" is the most critical problem faced by humanity, because good leadership is essential to surviving existential threats like climate change or nuclear war. With AI, it will mean the difference between utopia and dystopia.
posted by qxntpqbbbqxl at 9:19 AM on August 31, 2018 [8 favorites]

rarely centre the sole advantage we have as workers - the ability to withdraw our labour.

This, very much. What happens when the labor is withdrawn from the top? We're seeing a move towards mechanization in part to control labor costs and the rights of workers. Machines can't file Unfair Labor Practice complaints. Machines can't strike.

The article was both great and chilling. Well posted.
posted by corb at 10:02 AM on August 31, 2018 [3 favorites]

My natural inclination is to trust unvetted AI over anyone who uses the phrase "developing human consciousness" without irony.

The article is full of interesting questions and thoroughly unconvincing answers. If the alternative to being enslaved by bots that can successfully interpret Anna Karenina is embracing mystical ideas about human essence, it's going to be hard to decide which side to choose. I'm sorely tempted to throw myself on the mercy of the bots. But, I very much want to believe there's a third option. (Our material world is full of cows, but also raccoons, pigeons, geese, and cockroaches.)

Which isn't to say that preventing concentrated data ownership is a bad idea. How doing so provides meaningful jobs for people one hundred years from now isn't entirely clear. (That people one hundred years from now will find the idea of a job meaningful is also not entirely clear.)
posted by eotvos at 12:16 PM on August 31, 2018 [2 favorites]