Stephen Downes

Knowledge, Learning, Community

This is some clever commentary based on a clever AI prompt. The commentator jenka writes, "In the same way that English language emotion concepts have colonized psychology, AI dominated by American-influenced image sources is producing a new visual monoculture of facial expressions." To be fair, the actual prompt instructs the AI that "Everyone smiling directly at the camera." And the photos of stern-faced people jenka uses as comparisons "too come with an imposed, outside perspective." So what do we make of it? Well, to me, what we have here is a question related to the semantics of subjunctive conditionals. There are no real photos of, say, Aztec warriors taking a selfie, so what would such a photo look like? It depends (perhaps) on what you think the closest relevant possible world looks like. Are all Aztec warriors stern-faced? There's no way to know. But the vast majority of selfies depict people smiling broad American smiles. What else would the AI do? Why would you suppose it should do anything differently? Related: Evgeny Morozov draws a number of similar conclusions about what an AI would or should do, with no real consideration of what would make these claims true or false.

Today: 0 Total: 135 [Direct link]


Stephen Downes Stephen Downes, Casselman, Canada

Copyright 2023
Last Updated: Jun 02, 2023 5:42 p.m.

Canadian Flag Creative Commons License.