A mother and daughter are advocating for better protections for victims after AI-generated nude images of the teen and others were circulating.

A mother and her 14-year-old daughter are advocating for better protections for victims after AI-generated nude images of the teen and other female classmates were circulated at a high school in New Jersey.

Meanwhile, on the other side of the country, officials are investigating an incident involving a teenage boy who allegedly used artificial intelligence to create and distribute similar images of other students – also teen girls - that attend a high school in suburban Seattle, Washington.

The disturbing cases have put a spotlight yet again on explicit AI-generated material that overwhelmingly harms women and children and is booming online at an unprecedented rate. According to an analysis by independent researcher Genevieve Oh that was shared with The Associated Press, more than 143,000 new deepfake videos were posted online this year, which surpasses every other year combined.

  • virock@lemmy.world
    243·
    11 months ago

    I studied Computer Science so I know that the only way to teach an AI agent to stop drawing naked girls is to… give it pictures of naked girls so it can learn what not to draw :(

    • rustydomino@lemmy.worldEnglish
      2·
      11 months ago

      hmmm - I wonder it makes sense to use generative AI to create negative training data for things like CP. That would essentially be a victimless way to train the AIs. Of course, that creates the conundrum of who actually verifies the AI-generated training data…

      • gohixo9650@discuss.tchncs.de
        2·
        11 months ago

        this doesn’t work. AI still needs to know what is CP in order to create CP for negative use. So you need to first feed it with CP. Recent example of how OpenAI was labelling “bad text”

        The premise was simple: feed an AI with labeled examples of violence, hate speech, and sexual abuse, and that tool could learn to detect those forms of toxicity in the wild. That detector would be built into ChatGPT to check whether it was echoing the toxicity of its training data, and filter it out before it ever reached the user. It could also help scrub toxic text from the training datasets of future AI models.

        To get those labels, OpenAI sent tens of thousands of snippets of text to an outsourcing firm in Kenya, beginning in November 2021. Much of that text appeared to have been pulled from the darkest recesses of the internet. Some of it described situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self harm, and incest.

        source: https://time.com/6247678/openai-chatgpt-kenya-workers/