brainsteam.co.uk/brainsteam/content/annotations/2022/12/13/1670913121.md

date	hypothesis-meta	in-reply-to	tags	type	url
2022-12-13T06:32:01	created document flagged group hidden id links permissions tags target text updated uri user user_info 2022-12-13T06:32:01.500506+00:00 title The viral AI avatar app Lensa undressed me—without my consent false __world__ false 2iVhJnqvEe2HRauIjYpzBw html incontext json https://hypothes.is/a/2iVhJnqvEe2HRauIjYpzBw https://hyp.is/2iVhJnqvEe2HRauIjYpzBw/www.technologyreview.com/2022/12/12/1064751/the-viral-ai-avatar-app-lensa-undressed-me-without-my-consent/ https://hypothes.is/api/annotations/2iVhJnqvEe2HRauIjYpzBw admin delete read update acct:ravenscroftj@hypothes.is acct:ravenscroftj@hypothes.is group:__world__ acct:ravenscroftj@hypothes.is ml bias selector source endContainer endOffset startContainer startOffset type /div[1]/div[1]/main[1]/div[1]/div[2]/div[1]/div[1]/div[1]/div[2]/div[1]/div[1]/div[6]/div[1]/p[6] 245 /div[1]/div[1]/main[1]/div[1]/div[2]/div[1]/div[1]/div[1]/div[2]/div[1]/div[1]/div[6]/div[1]/p[6] 0 RangeSelector end start type 3237 2992 TextPositionSelector exact prefix suffix type AI training data is filled with racist stereotypes, pornography, and explicit images of rape, researchers Abeba Birhane, Vinay Uday Prabhu, and Emmanuel Kahembwe found after analyzing a data set similar to the one used to build Stable Diffusion. n historically disadvantaged.  It’s notable that their finding TextQuoteSelector https://www.technologyreview.com/2022/12/12/1064751/the-viral-ai-avatar-app-lensa-undressed-me-without-my-consent/ That is horrifying. You'd think that authors would attempt to remove or filter this kind of material. There are, after all models out there that are trained to find it. It makes me wonder what awful stuff is in the GPT-3 dataset too. 2022-12-13T06:43:06.391962+00:00 https://www.technologyreview.com/2022/12/12/1064751/the-viral-ai-avatar-app-lensa-undressed-me-without-my-consent/ acct:ravenscroftj@hypothes.is display_name James Ravenscroft	https://www.technologyreview.com/2022/12/12/1064751/the-viral-ai-avatar-app-lensa-undressed-me-without-my-consent/	ml bias hypothesis	annotation	/annotations/2022/12/13/1670913121

AI training data is filled with racist stereotypes, pornography, and explicit images of rape, researchers Abeba Birhane, Vinay Uday Prabhu, and Emmanuel Kahembwe found after analyzing a data set similar to the one used to build Stable Diffusion.

That is horrifying. You'd think that authors would attempt to remove or filter this kind of material. There are, after all models out there that are trained to find it. It makes me wonder what awful stuff is in the GPT-3 dataset too.