Post #2755612
2026-05-12 05:59 UTC
Replies (2)
-
@quincy@chaos.social 2026-05-12 06:01
@ErikJonker@mastodon.social @ElenLeFoll@fediscience.org But why should they? And why should that be "the future" and "unavoidable"? "Talking" to a bad representation of prior work instead of talking to real people, and doing the literature work properly instead of "talking" to a machine that invents some references and misses others? I don't see the point.
-
@mrosin@social.tchncs.de 2026-05-12 07:28
@ErikJonker@mastodon.social @ElenLeFoll@fediscience.org I'd argue yes, because it's unethical. Recently saw a presentation which can be summarized as “ethical AI is an oxymoron”: massive resource use, exploitative labor conditions, close ties of the industry to the ongoing global authoritarian resurgence, unauthorized and uncompensated use of other people's art, music, and literature, poisoning of the common information ecosystem (see above), de-skilling, disrupting communities with hyperscale data centers, etc. (“We are past the stage where we can say it is evil”... Are we? What has changed?) We don't reject unethical research practices because they produce bad results, even though they often do. We reject them because they're unethical. (Also, technological determinism is largely a discredited position.)