I just got thinking about the legislation around AI, and forecasting what might happen with the legalities around image-subjects not consenting for their image to be used, in ways they didn't actively opt into.
There's obvious arguments around "deepfake" videos/images, to prevent someone's likeness being widely distributed, and used to denigrate or compromise them in some way.
We currently think of this consent mostly in terms of "sexual" depictions. At present, we tend to gauge legality of any imagery created, based on whether the actual content is legal (eg. light nudity vs illegal acts). But, if laws were created around opt-in consent, you could apply this to virtually any imagery. A famous actor didn't consent to being "cast" in an AI video of him drinking wine, because he is teetotal or religious (for example).
So, I just got thinking about how future laws might shape up. Whether they would only apply to work that's publicly distributed (eg. on social media) or also private AI stuff created online (eg. Grok) if authorities demand their content be handed over for AI to trawl, for "spicy" words, images, etc. Like any emerging tech, lots of people are having lots of fun making nekkid pictures right now. Just like they did with early photography, videotape, and dial-up internet. But, if laws were to develop - and start operating retrospectively - could lots of people find themselves falling foul of that, for reasons going way beyond just erotic content? As in, virtually any kind of AI fake that draws a complaint from it's subject (or their lawyers)?
[link] [comments]