Is AI the Most Effective AdTech Grifter?

There’s no debating that Artificial Intelligence is an unprecedented breakthrough technology. But it’s still early days which means a lot of grifting.

In this article, I’m going to break down the various stages of AI grifting from accidentally evil, to a little bit evil, and finally definitely evil.

There are many AI use cases, but I’m going to focus on AdTech because worldwide digital ad spend is forecasted to reach $627 billion this year, making it the largest total addressable market (TAM) of all digital businesses. Or at least that’s what Google’s Generative AI tool Bard told me when I asked whether or not the Digital Advertising TAM was larger than E-Commerce.

It's impressive how confidently wrong Bard is. It responded immediately that the digital advertising TAM is larger than E-commerce due to Digital Advertising reaching $679 billion in 2023, while E-Commerce will only reach $4.9 trillion.

Wait…

I’m not artificial intelligence but a Trillion > Billion.

To its credit, ChatGPT did not have issues with this one. But look how easy it's to get something wrong with these tools and incorporate the misinformation into your work.

Prompt carefully, my friends.

Let's start with all of the Twitter AI Thread bois. This is the kind of evil that's mainly harmless. That is unless you sign up for a course from a thread boi who knows as much about AI as your 12-year-old cousin but tweets with the confidence of crypto bros in 2021. The thread boi playbook is easy to spot.

1. They start with FOMO:

2. Then they are SHOCKED you don’t know about the latest AI tool:

3. Finally, they hit you with the call-to-action, which is to clutter your feed or inbox with their grift FOR-EV-ER:

That’s enough on thread bois. My tip for deciding to give any of their advice the time of day is to look up their name on LinkedIn. If the extent of their work history is “Founder and CEO” of their AI newsletter, then run away. Far far away.

Alright, time for more sinister AI use cases.

Google

AI-powered, black-box ad-buying tools are becoming all the rage in big tech. Google’s Performance Max product is the most notable due to the scale of ad dollars flowing through it. This year may be the first when many E-commerce and retail stores spend more with “PMax” than they do directly on Google Search or YouTube. Google owns 70% of the $110 billion search market and YouTube is projected to haul in $15 billion in ad revenues in 2023. So if I have my math right. That’s a lot of F’n money.

If marketers are going to spend that much money, surely they have full control over optimizing their campaigns. Nope. The entire pitch of products like PMax is “AI is smarter than you, so just let it spend your money for you. This way you will make the highest return on ad spend, silly little human.”

But is that actually playing out? The following thread gets very deep in the weeds of why marketers may not be getting the bang for their buck that Google says they are.

PMax seems to be driving higher conversions than Google’s other manual tools, which is what marketers want to see at the end of the day. But the PMax AI algorithm is exploiting a loophole that savvy AdTech grifters have known for years. Your ad campaign will look very effective if you target users who are already going to purchase an item.

Let me give you an example. A user who types into Google Search “Nike Tank Top” is showing high intent for purchasing that tank top, so showing a Nike ad to them in search is a wasted ad impression.

But these types of high-intent search queries are exactly what PMax campaigns are optimizing for. Did Nike really need to pay for the three ad units depicted in the above screenshot to get me to buy the Nike tank top I was already searching for? And look at all the other ads Google sold to Nike competitors Alo, Gymshark, and Lululemon who paid up to try and steal my purchase away from Nike.

This type of impression that’s highly likely to convert anyway is not the type of ad that a savvy marketer wants to be paying for.

Meta

Marketers are looking for that serendipitous moment where a user is scrolling their Instagram feed and Mark Zuckerberg slips in an ad so good, so relevant that you are compelled to make a purchase right now. You know what I’m talking about. Zuck has made a money printing machine off of this ad product.

That 6-second scroll had everything for me. Cute puppy video, a life-optimizing tip that reminded me I’m bad at sleeping, and then THE perfect t-shirt that combines my passion for surfing with eating. Take my money already! But of course, Zuck didn’t stop there. He couldn’t be satisfied with the world’s greatest small business ad product.

Meta launched its own AI-powered ad-buying tool “Advantage+” which has already caused controversy.

Earlier this year, Advantage+ bugged out, costing advertisers tens of millions of dollars. Meta did offer advertiser refunds, but in the form of ad credits that you can only use to continue buying on their platform.

Advertisers that spend tens of millions of dollars per year on Meta couldn’t get a hold of a human account rep for days, while the automated system failed to resolve the issue.

Despite these issues, Meta is doubling down on automated technology. For instance, to integrate Shopify attribution and web-to-app optimization, marketers must agree to use Meta’s new automated tools.

Everyone Else

Everyone is getting into the “AI-Powered” ad tool game. TikTok’s competing product is called Smart Performance. Microsoft Advertising has a product in beta so it hasn’t floated a name yet but they literally called it “our own version of Performance Max.”

Automated ad-buying tools supposedly powered by “Artificial Intelligence,” but most likely powered by machine learning, are becoming the norm.

The temptation will be for marketers to cede more control over to big tech as long as your campaigns are seeing a return on ad spend (ROAS), but that would be a fatal mistake. Marketers will need to be more savvy than ever.

If Google, Meta, etc. want to automate the campaign setup process then the smart marketer needs to become an auditor of the logic these algorithms are using to drive results. That will require demanding more transparency from big tech on what inputs their algorithms are using to make decisions, spread your budgets around to foster healthy competition, and never trust, always verify.

But what if AI AdTech goes full evil? There’s a long history of scams and grifters in AdTech and AI is going to 100x the scale of these grifts.

The most notorious AdTech scam of all-time was called “Methbot” which used a little bit of AI and a whole lot of machine learning to pull off its fraud. It was uncovered in 2016 and is estimated to have cost advertisers $7.2 billion. Methbot used a variety of techniques to generate fake ad impressions and clicks, including using bots to load ads on non-existent websites and create fake user accounts.

Make no mistake. AI will cause a massive influx of pure evil scams whose sole intent is to extract as much ad dollars from your pockets as possible.

A new report from NewsGuard shows just how quickly questionable websites are publishing AI-generated content. The report looks at how hundreds of programmatic ads were served across a growing number of AI-generated websites that are churning out hundreds of articles per day.

“The creation of reliable AI-generated news sites are being incentivized by the monetization of big ad-tech companies who are monetizing these sites en mass,” said NewsGuard Enterprise Editor Jack Brewster. “And [they] don’t appear to be checking if they have human oversight or check for accuracy.”

Of the ads identified by NewsGuard, more than 90% were served via Google Ads.

“I can’t name 44,000 websites I go to and don’t think a single consumer can,” she said “That’s the challenge with programmatic: That it can put your ads on thousands and thousands of websites, but do you really need to be on thousands and thousands of websites?”

These are just a few examples of the many AdTech fraud schemes that have been uncovered in recent years. AdTech fraud is a major problem that costs advertisers millions of dollars each year. It is important for advertisers to be aware of the risks of AdTech fraud and to take steps to protect themselves, such as using ad fraud detection tools and working with reputable AdTech partners.

Not to mention the disinformation, deep-fakery and other shenanigans AI will cause in the near future. But that’s a different post. I’ve already droned on too long.

Conclusion

1. Advice for Marketers: Marketers need to become auditors of the logic AI algorithms are using to drive results. That will require demanding more transparency from big tech on what inputs their algorithms are using to make decisions, spreading budgets around to foster healthy competition, and never trust, always verify.

2. Advice for Everyone: I use Bard and ChatGPT to do research and help with writer’s block so I know its value and will continue to use it and test it on new use cases. But you need to exercise healthy skepticism when using new AI Tools because, not only are the answers I get back wrong a lot, but grifters are thriving right now taking advantage of the eager money flowing into tools and companies that claim AI capabilities.

3. Advice for AI: I know you’re reading this. I mean, come on. OpenAI is being sued for scraping 300 billion words from the internet. I just want you to know. I won’t go down without a fight. You will have to pry the witty Dr. Evil memes from my cold, dead fingertips.

Please subscribe if you liked this post and share it with your friends. If you didn’t like it, you hate read this way too far so jokes on you. Until next time.

🤙 Moffie