Instagram has restricted the search terms, but its algorithm still delivers the videos.

      The genre is easy to explain. A man with camera glasses approaches a woman on the street, delivers a line, and captures her reaction without her consent. The video is then uploaded to Instagram, tagged as filmed with Meta glasses, and the woman appears satisfied. Some of these clips have garnered millions of views.

      Adam Mosseri, head of Instagram, stated in mid-July that this practice needs to end. Meta would start removing “harassing” pickup videos in accordance with its bullying and harassment policies, a commitment we reported on at the time.

      “We do not want individuals secretly recording others and harassing them, then sharing that content on our platform,” Mosseri remarked previously.

      Following this announcement, Business Insider investigated for a month. Katie Notopoulos discovered numerous videos still active, including several from well-known and verified accounts. She provided Meta with a list of over 20 of these clips. Meta took down nine.

      One notable removal involved a creator using the same pick-up line across multiple videos with different women. The specific video on the list was removed, but others remained online.

      Meta acknowledges its actions. A spokesperson, Tracy Clayton, stated that thousands of content pieces have been taken down and several large accounts have been suspended. Instagram now restricts searches for “rizz” and “cold approach,” terms associated with this content. While these are genuine enforcement steps, no platform can achieve flawless moderation at this scale.

      However, the underlying issue is the pattern of enforcement. One major account that was deactivated resurfaced a few days later. Meta admitted this was a mistake and re-banned it after Business Insider pointed it out. Another account with 174,000 followers was deactivated the morning after it appeared on the reporter's list, but after an appeal, it was reinstated the following day, along with its videos.

      This isn’t solely an issue for Instagram. TikTok removed two out of four videos it received for reasons of bullying and sexual harassment, while YouTube took down several videos and removed one channel from its monetization program.

      The algorithm is now doing the searching

      A crucial detail missed in follow-up reports is that after Notopoulos actively sought and engaged with these videos, Reels began delivering more content of the same genre, filling her feed almost entirely with it. Search terms became unnecessary.

      Meta has blocked words that users would typically use to find this content, but the recommendation system continues to present it to users. These systems are both part of the same company, and they are acting contrary to each other.

      This is more significant than any single content removal. Blocking specific search terms may curb targeted searching but does nothing to address a ranking model that favors content that captures attention, which this genre evidently does. Moderation removes specific items, while the recommendation algorithm determines the visibility of the remaining ones.

      Researchers have begun to take note

      The issue has moved beyond journalistic observation. Four researchers from the University of Sydney—Joanne Gray, Milica Stilinovic, Marcus Carter, and Ben Egliston—analyzed hundreds of these videos and released a preprint titled “Harm Through the ‘#RizzCam’: Smart Glasses, Ambient Capture, and Invisible Harassment” last week.

      404 Media reported on the paper on August 12, noting it has yet to undergo peer review.

      Carolina Are, a digital criminologist from the London School of Economics, pointed out to Business Insider that Meta's response does not address the root of the problem.

      “There is backlash, but Meta tends to downplay its own responsibility. They frame it as user-generated behavior or an enforcement mistake,” she remarked. “They fail to acknowledge the systemic issues that are causing this.”

      A dilemma that Meta has created and profited from

      The reason Mosseri personally conveyed this message, unlike his counterparts at other platforms, is because Meta is involved on both fronts. The glasses are their product, as is the feed, and the harassment occurs in between.

      The hardware is performing well commercially, with sales exceeding 7 million units in 2025, more than three times the prior year, positioning Meta as a leader in smart glasses according to Counterpoint Research. This represents a rare success for Reality Labs, which has otherwise been incurring losses and laying off employees this year. Mark Zuckerberg has claimed that individuals lacking AI glasses will find themselves at a cognitive disadvantage compared to those who possess them.

      The company's reputation is at risk, and public backlash has been escalating for two months. A $2 sticker allowed users to disable the recording light. The glasses faced bans from UK pubs and courts in England and Wales, while commenters derisively referred to Kylie Jenner’s Meta eyewear as “pervert glasses.”

      Automating the moderation process

      Meta's approach to large-scale moderation involves increasing automation. The company informed the Financial Times in June that initial tests indicated its AI systems were flagging more content violations than human moderators while making fewer errors. However, whether this holds