Just had a thought-provoking lunch with Byron Reese and the topic of his new book is AI (releasing in April). I asked him, because they seem to be used interchangeably, is there a difference between AI and algorithms, because I have more of an issue with the latter than the former—in concept at least. My thinking runs to a working definition of AI that involves its ability to create revolutionary new things, not just evolve old things into the next generation things. Would either ever have put bacon on the donut maple log?
Now I’ll admit upfront that I’m a fan of the cautionary tale told in Cathy O’Neill’s “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy” that we are all too willing to assign authority to equations whose logical structure and assumptions we know very little about. But those are algorithms right, not AI. Or are they?
Byron pointed out that there is no really good definition of AI and that at its most rudimentary instance, algorithms can be considered AI. Well, that brings me back to bacons and donuts because what goes in, the inherent biases good and bad of the programmers determines the outcomes. As a case in point, look at the difficulty some facial recognition programs have with different races (e.g. they’re significantly more accurate with Caucasians).
So what does this have to do with marketing? A lot. And since AdTech and programmatic and computer-generated “creative” are some of my flash points—as well as catchall phrases for behavior that is a lot like the industry’s previous flight it and forget it mentality. Once an algorithm is blessed, there is little sense that it course corrects other than getting better than here it started.
As marketing becomes omnipresent (for better or for worse), actually connecting in a meaningful way with customers gets harder and harder. Think of NASCAR: How many logos does anyone actually see? Connecting brands and customers is no longer just about being there it is about recognizing each other as likely partners. And we all know how well the dating site algorithms do at predicting THAT little chemistry challenge.
“Robots and artificial intelligence will change the world,” Byron says, “empowering humans to be more productive and live better lives. We will use these technologies to end disease, hunger, and poverty.” In other words, they should be tools to better life, to allow more time for thinking, feeling and creating revolutionary solutions rather than evolutions. In marketing, as in life, we should be using our new tools rather than letting ourselves be dictated to by them.
Hear that HAL?