Recently, I read that as more people (myself included) have moved from using search engines to using AI to get answers instead, advertisers are going to start buying ad space targeted to AI assistants to prompt them to recommend their products to the user who asked. My first thought was “how long until the scammers start to show up and people start trying to game it?”
The Internet
There was once a time when the Internet seemed like a friendly and safe place. I used to get away with a four character password. Then the spammers, scammers, and hackers showed up, making it less safe. The trolls showed up, making it less friendly. Our email accounts were deluged with spam, setting off an arms race of spam and counter-offensives that lives on today.
Sure, there was plenty of BS before this, like chain letters, but mostly it was naive people who meant well and thought Bill Gates really would pay them millions of dollars to help him test an email tracking program by forwarding it to every single email they could get their hands on. Then there were the urban legends that people kept passing around. This is how Snopes came to life, and I used it heavily back in those days. I don’t even know if it’s still around. (I checked; it is.)
Search Engines
The purpose of search engines is to help you find things. But people found ways to game them. They would include a whole bunch of tiny white text on a white background that the search engines could read but users wouldn’t, which would put their post in the returns even if it had nothing to do with what the user was searching for.
Ebay
I haven’t touched Ebay since around 2007 when the iPhone came out. I was looking for a used iPhone, but when I searched for it, all the results came back for different phones that I didn’t want because people would include the word “not” in the title along with a list of phones people would be searching for. Every single item was titled something like “Flip phone NOT iPhone, Galaxy, Motorola, etc.” I don’t know if Ebay ever fixed this because I stopped using the site at that point.
Social Media
First, there was MySpace. It was a mess. Pages would scroll sideways for miles and had background music and it was impossible to find the play button to shut it off. But it allowed us to find each other. When I graduated high school in the early 90’s, you had to write letters to people. If you didn’t have their address and moved away, chances were you’d never see them again. Long distance calls cost a fortune so we avoided having to call people unless it was important. Same for military buddies. All of a sudden, people I assumed I’d never see again were finding me on MySpace and reconnecting. Some are still in touch.
Then came Facebook. Same thing. I managed to find some of the people I knew from MySpace and even more friends from my past.
I seem to remember Twitter (now X) coming along before Facebook, but that’s not right. In the beginning, Twitter’s premise was “In 140 characters or less, what are you doing right now?” It was fun in those days. There were friendly and interesting people.
Then the “social media consultants” showed up. All of a sudden, everybody you followed back immediately hit you with a DM trying to sell you services. Microsoft bought LinkedIn and seems to have turned it over to spammers and scammers. I hate LinkedIn and rarely go there. Every time I accept a connection request, somebody tries to sell me coaching, training, courses, or other things I am not interested in buying from random strangers on the Internet.
The Point
The point I’m trying to make is, I’ve lived through quite a few technology cycles and I’ve got the pattern down. When things are new and the pioneers are using them, they’re pretty cool. When the douchebags show up, that’s when things get rough.
And I’m pretty sure this “advertising to AI assistants” will follow a similar path. It could result in prompt injection, which essentially hijacks your prompt and has the AI do something entirely different. Or recommend things to you like “you should buy crypto!”
At that point, a trusted intelligence will become far more important than a super intelligence. As these AIs grow more integrated into our lives, we’re going to have to be able to trust them to look out for our best interests. We’re going to have to be able to trust the companies behind them to keep them secure from the developing threat landscape. And we’re going to have to be able to trust the people who sell the ad spaces to not sell them to people with evil intentions. Otherwise, the entire experience will get less pleasant.
As Joel Comm says, you are using the worse AI you ever will. 10 years from now, high school students may not believe that we had to type our prompts and read the response. But as the AI improves, so will the threats against it. We have to remain conscious of that.
What It May Look Like
So what will it look like when the douchebags arrive in AI advertising? There will likely be some hackers. There will likely be some scammers, some foreign operatives, political operatives, and the like.
The indicator you’ll be able to see the clearest will be people suddenly producing YouTube videos selling a $997 course or ebook or consulting to “make millions advertising to AI assistants”.
