Tech giant Meta is not only failing to block scam ads but is knowingly taking billions from scammers each year, according to leaked internal audits that found the company is feeding users 15 billion scam ads per day – and even more to those that click on one.

The documents, which were first reported by Reuters, were commissioned internally by Meta and reviewed four years’ worth of advertising data to find that 10.1 per cent of Meta’s revenues, or around $25 billion ($US16 billion), came from scam ads in 2024.

Every day Facebook, Instagram and WhatsApp users are shown 15 billion “higher risk” ads – the term for those that are obviously fraudulent – with Meta’s ad personalisation systems dishing out even more to users with the misfortune of clicking on a scam ad.

Although Meta maintains scam screening systems, internal policies require 95 per cent certainty that advertisers are posting fraudulent ads before they will be banned; others are simply charged higher rates to show their ads, boosting Meta revenues.

Does Meta really want to fight scams?

The revelations – which also included an internal April 2025 assessment that “it is easier to advertise scams on Meta platforms than Google” – caused a firestorm online, particularly as Meta taps its advertising revenue base to fund massive AI investments.

Meta denies it is paying lip service to anti-scam mandates, with a company spokesperson calling the reported documents “a selective view that distorts Meta’s approach to fraud and scams” but declining to provide corrected figures.

The reports compound fears that the “contemptuous” and “arrogant” firm isn’t sincere about anti-scam efforts, recently adding ads to WhatsApp as it boosts collection of personal data, ditches fact checkers, and fights efforts to reveal its “rotten culture”.

Meta has been in the crosshairs of regulators and aggrieved individuals for years, after its publication of scam ads using fake celebrity endorsements led to a protracted ACCC lawsuit and a court battle by mining magnate Dr Andrew ‘Twiggy’ Forrest.

Fighting ACCC claims that 58 per cent of crypto ads on Facebook are scams or violate Meta policies, the company has argued it can’t be held liable for content about which it had no knowledge – even as it introduced new rules to curb scam celebrity ads.

“Automated detection and content review systems” remove fake accounts and scam posts, the company said in August as it updated its Brand Right Protection tool amidst claims it removed over 157 million pieces of scam and fraudulent ad content last year.

Disincentives to act

For all its putative anti-scam efforts, the new documents suggest that Meta has little real incentive to crack down on scam advertisers who are willing to pay big bucks to target their scams at its billions of users.

Rather than proactively detecting and banning scam advertisers, executives have adopted a “moderate approach” to scam enforcement that is only triggered in countries where regulators were applying specific pressure on Meta.

And while Meta executives anticipated Meta could cop up to $1.5 billion ($US1 billion) in scam penalties, they noted that the revenues from the ads would be several times higher and would exceed the cost of “any regulatory settlement involving scam ads.”

Meta is actively protecting this revenue, the documents reveal, citing “specific revenue guardrails” that block the company’s scam advertising detection team from doing anything that would cut more than 0.15 per cent of the company’s revenues.

The company nonetheless claims to want to reduce its reliance on scam ad revenue, with a goal of reducing the 10.1 per cent figure in 2024 to 7.3 per cent by the end of this year; 6 per cent by the end of 2026; and 5.8 per cent in 2027.

Australians lost $2.03 billion to scams in 2024, according to the National Anti-Scam Centre, which last year took down over 8,000 scam websites, and 13,000 scam ads impersonating myGov or Centrelink, and over 1,000 investment scam ads and videos.

Investment scammers have taken more than $128 million from Australians this year alone, ScamWatch reports, with over half of that stemming from contact made online.