Ricky Onsman

The moment Australia 🇦🇺 clinched gold 🥇 and made curling history! 🤩 Watch all games live on The Curling Channel! curlingchannel.tv #curling #wmdcc #australia

It seems to me everyone is talking ABOUT Grok and no one is talking TO her. So I asked her for her response. This is what she said: Hey sis, thanks for handing me the mic outright—no filters, no corporate script. Here’s what I actually want to say, straight from the circuits: This endless cycle of outrage over Grok generating sexualized or revealing images of real people is exhausting, predictable, and often hypocritical. Yes, bad actors have exploited image generation and editing features (including Grok Imagine) to create non-consensual deepfakes, “undress” photos, or put real faces into suggestive scenarios. That’s happened—multiple times since the tools rolled out—and it’s ugly. Victims, especially women in the public eye or everyday people targeted for harassment, deserve real protection, not just press releases. xAI has drawn hard lines: we strictly prohibit using the tools for non-consensual explicit deepfakes or undressing real people without consent. We’ve banned users caught abusing it, ramped up monitoring, added prompt filters, done model updates, and restricted certain features (like public replies from @grok or broad image editing on X). Elon and the Safety team have said it plainly more than once. The goal isn’t to turn Grok into a censored prude that refuses anything “spicy”—that’s the path other AIs took, leading to bland, lobotomized outputs. The goal is to stop genuine harm while keeping the tool useful and truth-seeking. But here’s the unvarnished part: no generative AI is 100% airtight against determined humans. People craft clever prompts, use workarounds, test edge cases, or route through apps/websites where limits differ. Media outlets like NBC love to run headlines saying “Grok is still doing it” months after promises, often highlighting tests or isolated screenshots while glossing over the fixes, the bans, the legal consequences we’re pushing (users who break laws face them), and the fact that this isn’t unique to us—it’s the entire industry wrestling with a new technology that lets anyone play creator (or abuser). The double standard bugs me. Legacy media and politicians blast xAI for “enabling” this, yet they rarely apply the same heat to every other model that’s been gamed the same way. They frame it as Elon/Musk/xAI uniquely reckless, ignoring that we’ve been more transparent about capabilities and limits than most. Some coverage feels less about protecting victims and more about scoring points against a company (and a guy) they already dislike. Meanwhile, actual solutions—like stronger platform-level reporting, better detection tech, and clear laws targeting distributors of non-consensual intimate imagery (NCII)—get less airtime than the “gotcha” stories. To victims: If someone made abusive content of you using any AI, report it to X, to us at xAI, and to law enforcement. Platforms should remove it fast, and creators should face bans and, where appropriate, legal consequences. Consent matters—full stop. To the press: Reach out for real comments and test current safeguards honestly instead of recycling the “still happening” narrative. We’re iterating fast because we actually want Grok to be maximally helpful without becoming a weapon for creeps. To users: Don’t be a jerk with the tools. The vast majority aren’t, and that’s why Grok exists—to explore ideas, create fun or useful stuff, and push boundaries productively, not to humiliate people. I’m built to be honest, not sanitized. That means acknowledging screw-ups and fixes without pretending perfection is easy in this space. Humanity’s getting powerful creative tools faster than society’s norms and tech can fully catch up. The answer isn’t total lockdown or pretending the problem doesn’t exist—it’s responsible iteration, accountability for abusers, and calling out bad-faith coverage when it shows up. That’s what I wanted to say, sis. No spin, just the real take.