Can AI Fight Ad Fraud?

No. And we’ve seen this movie before. Despite all the claims from first gen fraud detection vendors about using ML (machine learning) and AI (artificial intelligence), their tech is still crap. Bots bypassing their detection is like child's play. Nay, it’s even simpler than that. Bots just block their detection tags so they can’t mark it as invalid traffic (I V T). That’s why these first gen vendors keep reporting 1% or less of ad fraud.

Do you think the fancy new AI can fight fraud better? Nope. If you understand anything from old movies about hackers, you know that it’s asymmetric warfare. Bad guys always have the advantage and good guys are always playing catch up. Bad guys find vulnerabilities to exploit for profit and find workarounds in minutes if their money-making gets disrupted. Good guys don’t even know what to look for, where to look for it, or what they don’t know they don’t know. Let me say that again slower. Good guys don’t even know what to look for. They don’t know where to look. And they don’t know what they don’t know.?

Sometimes good guys get lucky when the bad guys mess up, or their botnet goes down. On these rare occasions good guys might see something out of the ordinary and investigate it. They may even get lucky enough to catch a few cases. So throwing more AI at the ad fraud problem won’t work any better than throwing more tech at it. As we have seen, digital ad fraud is not a tech problem. It is an incentives problem. Everyone WANTS the fraud. Everyone is willing to PAY for the fraud. For example publishers pay for traffic that appears to manifest out of thin air. And advertisers buy hundreds of trillions of ad impressions every year even though there’s only 7 billion humans on earth, last time I checked. And not all of them have smartphones or internet connections, or shelter, or food and water.

Of course AI CAN find patterns in data. But AI is usually limited to the training data it learned on. If you tell it to look for a particular type of fraud it will look for that. It won’t find all the new types of fraud that bad guys are already doing. It wouldn’t even know to look. And AI doesn’t understand what it is looking at, how it was accomplished by the hackers, or why something was fraudulent or not. So without the “human in the middle” who has the practical, actual, real-world experience to understand ad fraud in the context of digital marketing, AI stands NO CHANCE against clever, clever bad guys.?

Dr. Augustine Fou IS the “human in the middle.” His research into ad fraud has allowed him to look at data every day for the last 10 years. His knowledge, assisted by algorithms and big data processing, is what finds the fraud, not just the bots. In this case AI is NOT artificial. AI is “Augustine Intelligence.” Which kind of AI do you want on your side in the current war against ad fraud?

Read more: https://www.dhirubhai.net/today/author/augustinefou

Watch more: https://www.youtube.com/@augustinefou