Instagram and Facebook Keep Showing Drug Ads. Here's Why It's Hard to Stop.

Meta, the company that owns Instagram and Facebook, has a rule against selling drugs on its platforms. But researchers at the Tech Transparency Project recently found more than 450 ads selling real prescription painkillers like oxycodone and Percocet—and illegal drugs like cocaine. These ads made it through Meta's screening systems, both the computer programs that check for rule violations and the people who review ads for a living.

The ads were paid promotions, meaning they went through the standard advertising process on Meta's platforms. This reveals cracks in how Meta polices what gets sold through those ads.

Why This Is Harder Than It Looks

Meta processes millions of ads every single day across all its apps. The company has spent millions building computer systems designed to spot drug sales and other illegal content. But the new research shows those systems are missing a lot.

Meta says it takes the issue seriously and removes drug sales when it finds them. The company's official rules ban people from coordinating drug deals on the platform. Yet enforcement has been inconsistent.

The problem is not unique to Meta. Researchers at RAND Corporation found examples of synthetic opioids and the chemicals used to make them available for purchase across most major social media platforms. This suggests the issue is industry-wide, not just a Meta problem.

A Bigger Industry Response

Meta has joined groups working to stop synthetic drug sales online. Other big platforms like Snap are working on this too. The U.S. government and the United Nations are now coordinating with these companies to try to disrupt drug trafficking on social media.

This kind of coordination between platforms, law enforcement, and governments is becoming more common. We have seen this pattern before when e-commerce sites like eBay and Amazon dealt with counterfeit goods in the early 2000s. Those platforms eventually built better detection systems and started working with law enforcement. Drug sales on social media may follow a similar path.

The scale and nature of this problem has grown sharper with drugs like fentanyl. These synthetic drugs can be made with chemicals that are easy to get, and they spread through decentralized networks that are well-suited to how social media works.

How Drug Sellers Stay Hidden

Drug sellers use tricks to avoid detection. They rely on emojis that have coded meanings, use vague language, and hide their messages inside what look like normal pharmaceutical advertising. A computer program can easily mistake a coded drug ad for a legitimate ad for prescription medication.

When people buy ads on Meta, those ads get checked in multiple stages. A computer program looks at it first, then sometimes a person reviews it, and the platform keeps monitoring it after it goes live. Yet drug sellers still get through.

The real challenge is that drug sellers watch what gets removed and quickly change their tactics. It becomes an ongoing battle: the platform removes an ad, the seller uses different words or emojis, and the process starts over. Meanwhile, legitimate pharmaceutical companies are allowed to advertise prescription drugs, which creates gray areas that bad actors can exploit.

What This Means Looking Ahead

The fact that drug ads are showing up across multiple platforms at once suggests that the current way platforms police content may not be built for this problem. Platforms typically find and remove bad content after it has already been posted. That reactive approach moves too slowly against drug sellers who can adapt and change platforms instantly.

The newer approach—having platforms, law enforcement, and government agencies work together—may be more effective. But whether it will work remains to be seen. If it does succeed, similar coordinated efforts might be used to tackle other kinds of harmful content that spread across social media.

What this highlights is that keeping social media safe at the scale these platforms operate requires more than just automated detection systems or rules. It may require rethinking how platforms are built from the ground up to make coordinated harmful activity harder to hide in the first place.