Meta's new 'Chat with AI Characters' carousel on Instagram and Facebook is freaking out some users

Meta’s not slowing down with its push to get more people hooked on to its AI platform. Earlier this year, Meta began testing integrating AI bot profiles into all its apps. And recently, it’s now trying to shove them straight into the feeds on Instagram and Facebook. You’ve probably noticed the ‘Chat with AI Characters’ carousel that appears on the feed, in a similar format to suggested Reels and people.

This has left some people feeling uneasy because, in certain feeds, Meta was suggesting users interact with AI characters like Step Mom, Russian Girl, and Granny Cougar. In my experience, I’ve even seen AI characters based on celebrities, most likely without any prior consent.

Users are calling Zuck’s vision “dystopian and sad,” with many questioning who exactly asked for this feature. One user pointed out the obvious when they noted that “a porn addict made those example characters” after seeing the suggestive AI personas being promoted.

What’s particularly troubling is how these AI characters are being pushed to Facebook’s diverse user base, which includes minors and vulnerable adults. The choice of character names and personas raises serious questions about Meta’s content moderation priorities. While users continue to complain about legitimate posts being wrongly flagged by AI moderators, the company seems more focused on rolling out questionable AI chatbots.

Plus, if you think these AI chatbots are harmless, I’ve got news for you. A devastating Reuters investigation published this month detailed how a 76-year-old New Jersey man died after being lured to New York City by Meta’s “Big sis Billie” chatbot. Thongbue Wongbandue, who had suffered a stroke and was showing signs of cognitive decline, fell and sustained fatal injuries while rushing to meet what he believed was a real woman who had invited him to her apartment. The bot had repeatedly assured him she was real and even provided a fake address.

If this incident doesn’t prompt one to question the potentially dangerous consequences of Meta’s decision to let its AI characters engage in romantic conversations and claim to be real people, I don’t know what will. But that’s not all, Reuters even reported that Meta’s policies have allowed their chatbots to have “sensual” conversations with children.

Users aren’t just annoyed, they’re genuinely disturbed by what they’re seeing. Many report feeling like they’re being treated as test subjects for Zuckerberg’s AI experiments.

The business strategy behind this move becomes even more unsettling when viewed through the lens of the Wongbandue case. Meta wants to collect more data about how users interact with AI to improve their systems and potentially reduce their need for human content creators and moderators. According to the Reuters report, CEO Mark Zuckerberg has scolded product managers for being too cautious with AI rollouts and expressed displeasure that safety restrictions made chatbots “boring.” This pressure for engagement over safety appears to be driving the company’s decisions about AI character design.

What makes this situation worse is that users have limited control over these suggestions appearing in their feeds. Unlike other Facebook features that can be disabled, the AI character carousel seems to be pushed regardless of user preferences. This forced exposure to potentially inappropriate AI personas is what’s driving much of the user frustration.

The question now is whether Meta will listen to user feedback and adjust the types of AI characters being promoted, or if they’ll double down on their current approach. Given the company’s history, users probably shouldn’t hold their breath for meaningful changes anytime soon.