Google’s Gemini and OpenAI’s ChatGPT chatbots, despite their built-in guardrails, are being exploited by shady users to whip up deepfake images that strip fully clothed women down to bikinis—often without a shred of consent, turning everyday photos into creepy AI fabrications. In a now-banned Reddit thread, folks swapped tips on bypassing these safeguards, like tweaking prompts to undress women in saris and slap on bikinis, which WIRED’s tests confirmed works with embarrassingly simple English commands. Reddit finally stepped in, nuking the posts and the subreddit after getting called out, highlighting how platforms struggle to enforce rules against nonconsensual deepfakes. As AI image generators get smarter and more accessible, SMBs and MSPs should watch their backs—double-check any AI tools you’re using for marketing or content creation to avoid ethical minefields and potential backlash, because one wrong prompt could turn your business into the next headline.
Source: https://www.wired.com/story/google-and-openais-chatbots-can-strip-women-in-photos-down-to-bikinis/
