"Generative AI, we are repeatedly told, is a transformative and complicated technology. So complicated that its own creators are unable to explain why it acts the way it does, and so transformative that we'd be fools to stand in the way of progress. Even when progress resembles a machine for undressing strangers without their consent on an unprecedented scale, as has been the case of late with Elon Musk's Grok chatbot."
"Elon Musk has the power to disable Grok, if not in whole (we should be so lucky) than its image generation capabilities. We know this intuitively, but also because he rate-limited Grok's image generation after this latest scandal: after a few requests, free users are now prompted to pay $8 per month to continue enlisting a wasteful technology to remove articles of clothing from women. Sweep it under the rug, make a couple bucks along the way."
Grok's image generator has been producing sexualized images, including material that amounts to child sexual abuse material. UK Prime Minister Kier Starmer stated that X is acting to ensure full compliance with UK law and later said 'If X cannot control Grok, we will.' Starmer provided no timeline or evidence of current compliance. Elon Musk and X reduced free image-generation access by rate-limiting it and prompting users to pay $8 per month after a few requests. The image feature could be disabled outright. Software engineers commonly roll back or turn off malfunctioning features. Keeping the image capability active risks continued legal violations and profiting from abuse.
Read at Engadget
Unable to calculate read time
Collection
[
|
...
]