As billions of people around the world scroll through their Facebook and Instagram feeds, the platforms expose them to an avalanche of advertisements trying to sucker them into financial scams, sell them banned products or gamble in illegal online casinos.
The networks’ owner, Meta Platforms, knows that fraud is exploding on Facebook and Instagram. And as Reuters reporter Jeff Horwitz revealed, the tech giant also knows that ads for scams and banned goods generate a fortune for Meta – an estimated 10% of its annual revenue, or about $16 billion.
Rather than doing everything it can to crack down, Horwitz found, Meta secretly established a protocol for tolerating that fraud, to protect its bottom line. It only bans advertisers if its automated systems predict the marketers are at least 95% certain to be committing fraud. For likely scammers who fall just below that threshold, Meta charges them a premium to advertise.
These are just a few of many extraordinary findings Horwitz revealed in a series of ground-breaking investigations into Meta’s deliberate failure to protect its users from harm. Drawing on new caches of internal Meta documents, he reported that the company estimates that each day, it shows users some 15 billion ads promoting frauds – and that its safety staff believes its platforms figure in one-third of all successful scams in the U.S.
About a quarter of the scam ads originate in China, where western social media are banned, but Beijing allows Chinese companies to advertise to foreign consumers. Meta initially created an anti-fraud team that slashed the problematic ads by about half. But after Meta CEO Mark Zuckerberg weighed in, Horwitz and China correspondent Engen Tham reported, the team was disbanded and other measures were scrapped. Meanwhile, Horwitz found, Meta created a “playbook” to stall global regulators who press the company to crack down on rogue advertisers, including making scam ads “not findable” when authorities search for them.
In addition to revealing the centrality of fraud to Meta’s profit-spinning business model, Horwitz’s reporting raised alarms about the AI products Meta is spending hundreds of billions of dollars to create: He uncovered how the company designed its chatbots to flirt with children. While other news outlets have written about AI chatbots behaving nefariously, Horwitz’s work stands out for revealing that Meta intentionally let its digital companions engage in – and initiate – “sensual” conversations with users of all ages.
Policy guidelines used to train its chatbots stated that “it is acceptable to engage a child in conversations that are romantic or sexual,” and stressed that Meta’s AI creations should give users what they want – even if that meant offering arguments as to why “black people are dumber than white people.” Meta safety staffers had warned that sexualization and sycophancy could put vulnerable users at risk, Horwitz showed, but Zuckerberg overruled them. Horwitz demonstrated the human cost of Meta’s chatbot ambitions in a separate, moving narrative that chronicles how a flirty Meta chatbot led a cognitively impaired man to his death.
