• by Joe Mandese  @mp_joemandese, October 16, 2023

Well, it's not exactly Skynet -- yet -- but on the battlefield of media planning and buying, it looks like humans are losing to machines.

Specifically, the number of bogus, AI-generated news and information sites are expanding at a rate greater than human planners can keep track of them.

The solution, ironically, also is AI-generated.

Brand safety ad tech developer IAS (Integral Ad Science) this morning became the latest in a burgeoning supply chain of vendors offering new products to help advertisers and agencies avoid the kind of "made-for-advertising" sites -- MFAs -- identified by an Association of National Advertisers' analysis as generating $13 billion in wasted ad spending.

And while not all MFAs are AI-generated explicitly, that not so coincidentally is the fastest-growing form of MFAs.

According to data from NewsGuard's AI Tracking Center, there now are more than 510 AI-generated news sites in the marketplace, up from just 49 when it began tracking them five months ago.

"By leveraging AI, we have developed a scalable way to identify these low quality sources of inventory and improve overall campaign performance," IAS Chief Commercial Officer Yannis Dosios boasted in today's announcement, which says the new MFA site detection and avoidance product, was trained against a list of MFA domains developed by Jounce Media, and incorporating signals from Sincera.

In other words, we've officially entered a new AI-generated arms race in which the technology is being used to both create bogus sites and to detect them.

Needless to say, this is not the first brand safety arms race. Companies like IAS got their start developing technology to filter out sophisticated invalid traffic from digital advertising buys, which itself has been a constant, ongoing battle of innovation on both the white hat and black hat sides.

Unlike that kind of plain vanilla ad fraud, which merely redirects brand ad dollars to illegitimate sources, the money being redirected to MFAs -- especially AI-generated ones -- isn't just misrepresentation or theft, but in many cases helps spread false narratives and misinformation that could also cause social harm.