Skip to content
Pioneer Page PP markPPPIONEERPAGE

Pages from the advancing edge of computing.

Foundation models

Meta docs project $16B from scam ads; Cuban cites weekly deepfake victimization

Internal documents show Meta projected $16 billion in revenue from fraudulent advertisements, while Mark Cuban describes receiving weekly deepfake scam emails.

Meta docs project $16B from scam ads; Cuban cites weekly deepfake victimization
Depth
$16 billionReuters reported on November 6, 2025, that Meta expected roughly $16 billion from scam ads · reported
95%Meta's systems require 95% confidence to ban an advertiser, allowing lower-confidence accounts to remain live · reported
10 percentThis figure represented about 10 percent of Meta's total revenue for 2024 · reported
1 outlet · newest source 403h ago

Key points

  • Reuters reported on November 6, 2025, that Meta expected roughly $16 billion from scam ads.
  • Mark Cuban said he receives weekly emails about fake diabetes cures using his AI-generated likeness.
  • Meta's systems require 95% confidence to ban an advertiser, allowing lower-confidence accounts to remain live.

Reuters reported on November 6, 2025, that internal Meta documents projected the company would earn approximately $16 billion from advertisements promoting scams and prohibited products.reported This figure represented about 10 percent of Meta's total revenue for 2024.Internal documents show Meta expected to earn $16 billion from scam ads in 2024.Reuters reported Meta's internal projection of $16 billion in revenue from fraudulent advertisements in 2024. The documents indicated users saw roughly 15 billion "higher risk" ads daily across Meta's applications.

Evidence: reportedCuban: Facebook knowingly runs AI deepfake scam ads; docs say $16B fraud revenue

aifleet-hn-ai

Evidence: reportedCuban: Facebook knowingly runs AI deepfake scam ads; docs say $16B fraud revenue

aifleet-hn-ai

How we got here

Mark Cuban told listeners of the Pivot podcast in August that he receives weekly emails from elderly people asking about a $294 diabetes cure he never sells. He stated these ads use AI-generated copies of his face and voice. Cuban noted that some ads include disclaimers admitting they are AI-generated entertainment, but argued the fraudulent pitch succeeds before users read them. He called Meta "the worst of the worst" for allowing such content.

Evidence: reportedCuban: Facebook knowingly runs AI deepfake scam ads; docs say $16B fraud revenue

aifleet-hn-ai

What is new

A late 2024 internal document valued the "higher risk" ad category at about $7 billion annually. In a May 2025 presentation, Meta's safety staff estimated their platforms were involved in one-third of successful scams in the U.S. The documents revealed that Meta's automated systems must assign a 95% likelihood of fraud to an account before banning it.

Evidence: reportedCuban: Facebook knowingly runs AI deepfake scam ads; docs say $16B fraud revenue

aifleet-hn-ai

Reading between the lines

Cuban argued that the presence of these ads defines Meta's organizational culture. He suggested that a company tolerating scams targeting grandparents demonstrates what it will accept at every level. Cuban co-founded Cost Plus Drugs, an online pharmacy selling low-cost generic medications, which contrasts with the fraudulent cures advertised in his name.

The open questions

The source material presents Meta's internal projection of $16 billion as an expectation of revenue flow, not a stated corporate target or goal. Reuters reported this figure based on internal documents, but the extent to which Meta actively encouraged this revenue versus merely failing to prevent it remains unclear. The 95% threshold for banning advertisers means suspected scammers can stay live if the confidence level is lower, simply paying more per ad. It is unclear how many accounts remain live indefinitely below this threshold. Cuban's weekly emails are anecdotal evidence of the problem, but the material does not provide independent verification of the frequency or scale of his personal victimization beyond his own account.

Evidence: reportedCuban: Facebook knowingly runs AI deepfake scam ads; docs say $16B fraud revenue

aifleet-hn-ai

Evidence: reportedCuban: Facebook knowingly runs AI deepfake scam ads; docs say $16B fraud revenue

aifleet-hn-ai

Why it matters

  • Model evaluators must consider that automated content moderation systems may require extremely high confidence thresholds, such as 95%, before taking action.
  • The existence of "higher risk" ads suggests that Meta's classification systems allow certain fraudulent content to remain visible if it meets revenue criteria.
  • AI-generated deepfakes can include disclaimers that do not prevent user deception, challenging assumptions about transparency as a safety measure.
  • The scale of "higher risk" ads, estimated at 15 billion daily views, indicates significant exposure for users on these platforms.

Topics: Meta, deepfakes, fraud, AI safety, advertising

Updated

The strongest doubt

Counterpoint from the sources

Meta's safety staff estimated their role in successful scams, suggesting awareness of the issue, but the documents do not indicate a change in policy to lower the 95% ban threshold.

Numbers in context

95%reported10 percentreportedCircle area is proportional to the magnitude; labels retain signs. Figures may describe different measures. Read their contexts above.

Ask the article

Local keyword search of this article; no generated answers.

Did this article give you the evidence you needed?

No vote saved on this device.
Your choice stays on this device. No public tally.

Questions readers ask

How much revenue does Meta expect from scam ads?

Reuters reported that internal documents showed Meta expected roughly $16 billion from such ads in 2024.

What is the threshold for banning an advertiser?

Meta's automated systems must be 95% likely that an account is running a scam before banning it.

Did Mark Cuban sell the diabetes cure advertised in deepfakes?

No, Cuban stated he never sold the cure and receives weekly emails from victims.

Connected stories

This storyDemocratic lawmakers signal subpoena-driven AI and data center oversightDemocratic lawmakers signal subpHinton cites corporate race and trust gaps as primary AI risksHinton cites corporate race and Abliteration AI removes safeguards from open-weight models for developersAbliteration AI removes safeguareBay browsing by Meta Muse agents appears blocked per user reportseBay browsing by Meta Muse agent

Reading receipt

  • Reuters reported on November 6, 2025, that Meta expected roughly $16 billion from scam ads.
  • Mark Cuban said he receives weekly emails about fake diabetes cures using his AI-generated likeness.
  • Meta's systems require 95% confidence to ban an advertiser, allowing lower-confidence accounts to remain live.
Time spent: 0:00 · Depth: Full

Sources

  1. Hacker News: AI posts over 50 points

Related stories

Background reading