Techno-sociologist Zeynep Tufekci provides a stark view of the potential future of artificial intelligence (AI) and the possible dystopia toward which we are heading. As she explains, this is not just the next thing after online ads, but instead a major leap in category, as with AI, persuasion architectures can be built at the scale of billions of people, and then deployed at individuals. Examples she gives include targeting bipolar patients with ads about Las Vegas trips as they’re entering the manic phase of their condition, or repeated experiments by Facebook that showed how a small tweak in a message display led to hundreds of thousands of additional voters in elections. Because all of these algorithms and activities are hidden (she cites the Trump campaign’s use of paid, private, non-public posts meant to depress turnout from African American voters), there is incredible power here to control elections, all stemming from technology built to get people to click on ads. She calls instead for a future where AI supports us in our human goals but is constrained by our human values, and a digital economy where our data and attention are not for sale.
1 Thought on "Artificial Intelligence and the Infrastructure of Surveillance Authoritarianism"
Thanks for increasing awareness of this very important thinker. Tufekci is well worth following. (Maybe a keynote speaker one of these days at the SSP meeting?)