Dr. Augustine Fou says AI is scaling ad fraud industrially
Dr. Augustine Fou explains how Artificial Intelligence is contributing to the industrial-scale expansion of ad fraud, detailing practices like fabricated performance metrics and spoofed supply paths. The discussion highlights how current ad systems allow significant 'take rates' and markups, resulting in publishers receiving a fraction of the advertiser's bid. Fou suggests that increasing bids often benefits intermediaries rather than improving inventory quality.
Key Takeaways
- A $1.50 CPM can shrink to $0.15 for the publisher after intermediaries take their cut.
- Fou says increasing a bid often enriches the middleman instead of improving inventory quality.
- He cites fabricated performance data, including fake clicks, fake conversions, and fake analytics.
- The episode calls out spoofed supply paths as part of the fraud layer advertisers do not see.
Why It Matters
The immediate issue is that AI can make ad fraud cheaper and easier to scale, while publishers still receive only a small fraction of advertiser spend. The broader problem is structural: Fou describes a system with take rates, markups, and spoofed supply paths that lets intermediaries profit even when reported performance is fabricated. For StreamingMeme readers, the key signal to watch is whether ad-tech buyers and publishers start focusing on the CPM-to-publisher gap, like the $1.50-to-$0.15 example, instead of headline performance metrics.
Read full article at youtube.com