'Protecting its kind': Instagram’s AI is banning humans while giving verified mark to bots and scammers

Meta's AI is currently suffering from a massive capability gap, proving utterly incapable of telling the difference between real human beings and fraudulent accounts. If Meta doesn't abandon its current AI project and rely on more competent third-party LLMs for protection, Instagram is headed for the same slow irrelevance that killed MySpace, Orkut, and Facebook.

A massive moderation crisis is currently tearing through Instagram, exposing a glaring paradox at the heart of Meta’s social media empire. Across the globe, legitimate users are watching helplessly as their personal and business accounts are suddenly deactivated, while coordinated bot farms and scammers effortlessly secure the coveted "Verified" blue checkmark.

The root of this chaos lies in Meta’s aggressive reliance on artificial intelligence. In a sweeping effort to cut costs, the company has largely replaced human trust-and-safety moderators with automated algorithms. The official line from Meta routinely describes sudden dips in user metrics as "routine cleanups" designed to eliminate fake accounts. However, the reality experienced by millions tells a much darker story: the AI is fundamentally broken.

Real humans are finding themselves permanently banned for everyday behaviours. Liking too many posts, logging in from a new device, or simply commenting too quickly is now enough to trigger the algorithm’s "inauthentic behaviour" filters. Worse yet, because human support has been virtually eliminated, users attempting to appeal these unfair bans find their cases reviewed by the exact same flawed AI that punished them in the first place, creating an endless, frustrating loop of automated rejection. This is because of the mass layoffs that Meta has done in recent months.

Meanwhile, scammers are exploiting the system with frightening ease. Sophisticated bot networks, cryptocurrency fraudsters, and spam accounts are successfully navigating Meta’s automated verification process, purchasing subscriptions to secure the blue checkmark. This badge, once a symbol of authenticity and trust, is now frequently used to legitimise predatory behaviour. The AI system scrutinises the genuine human user base with extreme prejudice but remains entirely blind to these automated bad actors.

Tech analysts suggest this may be a symptom of "model collapse." With the platform saturated by automated traffic, the AI may be losing its baseline understanding of normal human interaction. If a real human doesn't perfectly match the AI's increasingly flawed definition of authenticity, they are purged.

Until Meta reintroduces genuine human oversight, the platform remains increasingly hostile to its actual users, punishing humanity while rewarding the bots.