Shitty behaviour around the img2dataset AI scraper
The author of this popular AI training data scraping tool doesn’t seem to understand consent and opt-in:
Letting a small minority [ie web publishers] prevent the large majority [AI users] from sharing their images and from having the benefit of last gen AI tool would definitely be unethical yes. Consent is obviously not unethical. You can give your consent for anything if you wish. It seems you’re trying to decide for million of other people without asking them for their consent.
In other words, “scraping your content without opt-in is better than denying access to your content for millions of potential future AI users”. An issue to implement robots.txt support has been languishing since 2021. Good arguments for blocking the img2dataset user agent in general…Why is British media so transphobic?
Aside from the weirdness of Mumsnet, I didn’t know about the influence of the mid-2000s skeptics movement:
While claiming to be the country’s foremost critical thinkers, the group was riddled with anti-humanities bias and a fetish for a certain kind of “science” that it held to reveal a set of immutable principles upon which the world was built with almost no regard whatsoever for interpretative analysis based on social or historical factors. Part of this mode of thinking was an especially reductivist biologism: the idea that there are immutable realities to be found in our DNA, and if we just paid enough attention to Science and stopped trying to split hairs and discover meaning over in the superfluous disciplines of the humanities, then everything would be much simpler. It’s precisely this kind of biological essentialism — which skirts dangerously close to eugenics — that leads people to think they can “debunk” a person’s claim to their gender identity, or that it should be subjected to rigorous testing by someone in a lab coat before we can believe the subject is who they say they are.
(tags: debunking scepticism skeptics history terfs uk uk-politics gender)