Scapegoating the Algorithm
asteriskmag.com
Scapegoating the Algorithm
1–10 of 36 posts
Re: Scapegoating the Algorithm
#2> Our research is dedicated to ensuring AI is a force for good and it’s structured in a series of research programmes that cover a wide range of projects. Our work explores vital questions about the risks and opportunities emerging with AI in the near, mid- and long-term. These range from algorithmic transparency and the nature of intelligence to automated warfare, consciousness, social AI, feminist AI, AI-amplified injustice, global and pluriversal design, and the implications of AI for democracy, geopolitics, and the natural environment.
Re: Scapegoating the Algorithm
#3Author is Associate Fellow at the Leverhulme Centre for the Future of Intelligence (CFI), https://www.lcfi.ac.uk/about > Our research is dedicated to ensuring AI is a force for good and it’s structured in a series of research programmes that cover a wide range of projects. Our work explores vital questions about the risks and opportunities emerging with AI in the near, mid- and long-term. These range from algorithmic…
Re: Scapegoating the Algorithm
#4Re: Scapegoating the Algorithm
#5But the real issues are not the ones to which he points. Our problem is that as a society we have no real values. Most people will take whatever job pays the most money irrespective of social consequences. Politicians will engage in arguably corrupt behavior as long as benefits outweigh costs.
Google, by pioneering an internet based on advertising, shifted the allocation of the collective Internet-capital toward producing content that is engaging. You tell me if engagement correlates with truthfulness or long term utility (it doesn't). Google -- though yes initially useful and even utility-generating -- would ultimately extract all the useful latent value (including any surplus it had added) from the internet. We are left with a very high entropy internet, where you are far more likely to find factually incorrect or simply worthless content (noise) than useful content (signal). (Though, yes, bastions of order do still exist, like wikipedia.)
Social media (facebook) simply did the same but instead of the broad internet, focused on extraction of value from interpersonal relationships. The result there, too, has not been good.
So is the internet / social media different? Yes and no. The degree of concentration of power, potential for manipulation, and general capacity to shape the world is much greater in these companies. Whatever trends existed previously have been substantially accelerated. These companies have means to influence nearly every aspect of life. That is not something that magazines, radio, nor (untargeted) television could accomplish.
Moreover, given their substantial power, companies like Goog and FB are more capable of altering the fabric of values that might otherwise have resisted change; i.e., they have accelerated the decline of institutions that would have otherwise favored truth, community, etc.
America's "large scale epistemic challenges" are exactly that: our society and institutions are increasingly devoid of concern for what truth is. There aren't "right" answers to every problem, but to have a debate, there has to be some set of values against which to measure consequences, and good faith commitment to a framework to measure. That's a notion of "truth", and we have mostly lost that.
Social media makes zero epistemic commitments, except whether a marginal dollar is earned. Though it is not the only problem, if your society were overrun by drug dealers turning people into mindless zombies, you might realize that it's hard to fix anything until you expel the drug dealers.
This author -- at the Leverhulme Center for the Future of Intelligence -- likely wants to use algorithms for democracy; i.e., his group has the new miracle pill to fix your ill.
Re: Scapegoating the Algorithm
#6Things seem a lot dumber since social media, but I guess it is possible that the same dumbness is just being broadcast wider now that there aren’t any gatekeepers, I guess.
Re: Scapegoating the Algorithm
#7Ugh, this fellow misses the forest for the trees. Though he's partially right: social media is partially an extension of past trends. See, e.g., "Amusing Ourselves to Death" by Postman. But the real issues are not the ones to which he points. Our problem is that as a society we have no real values. Most people will take whatever job pays the most money irrespective of social consequences. Politicians will engage in a…
Maybe if there was a social safety net, people would feel more comfortable choosing options that pay less. My soul has a pretty low price since I value security over abstract morality.
Re: Scapegoating the Algorithm
#8Ugh, this fellow misses the forest for the trees. Though he's partially right: social media is partially an extension of past trends. See, e.g., "Amusing Ourselves to Death" by Postman. But the real issues are not the ones to which he points. Our problem is that as a society we have no real values. Most people will take whatever job pays the most money irrespective of social consequences. Politicians will engage in a…
> Most people will take whatever job pays the most money irrespective of social consequences Maybe if there was a social safety net, people would feel more comfortable choosing options that pay less. My soul has a pretty low price since I value security over abstract morality.
My main objective is not to critique that decision, but rather to argue that a society dedicated to the pursuit of dollars is likely to find itself epistemically astray
Re: Scapegoating the Algorithm
#9Re: Scapegoating the Algorithm
#10The author is wrong about psychology: people are generally not savvy information consumers. They mostly converge on the average of what they see around them. Cult leaders use this to their advantage by removing people from family and non-biased sources of information. The human brain acclimates and it's hard to break away from that situation epistemically.
Advertising generally works and is well measured. The process of selling people Coke or Pepsi is not fundamentally different from selling them on political ideas. And in practice many leaders have found it to be of practical utility to strengthen their power with a socially promoted ideology, whether that's religion in ancient times or state religion during the Soviet era or conspiracy theories in the current era.
I'd like to see people who are skeptical of the power of propaganda tackle these issues. They tend to cite a handful of reports claiming that propaganda was ineffective in 2016, but those reports were not well done and some members of the intelligence community have publicly stated that foreign influence was decisive in the 2016 election. The official reports that I'm aware of deliberately made no assessment of the impact on the election results.
If one believes that such influence is not effective, then one would have a harder time explaining why we're seeing more countries copy the Russian model. Clearly their militaries believe that it is effective. And one would also have a hard time explaining why the US engages in similar tactics abroad, including promoting anti-vax content in China.
Anyway, I see why people make the sort of argument the author is making. But it doesn't seem psychologically plausible or empirically correct. And it spreads the meme that consuming propaganda 24 hours a day isn't bad for you