We've long known that AI can help algorithms make more precise recommendations online. If you've bought new pots and pans, Amazon may recommend new kitchen knives to complement your purchase, since many shoppers replacing pots and pans are also upgrading their kitchen tools. If you've browsed some new business attire at one store, car companies might target you with ads for new vehicles, assuming that people refreshing their work wardrobes are often in a position to upgrade their cars. If you've watched a cute cat video on TikTok, expect to see more — because let's face it, there's no such thing as too many cat videos.

But what if the algorithm can do more than just figure out your most likely behaviour based on other people's past actions? What if our digital footprint is so robust, and AI is getting so sophisticated, it can make predictions based on our unique personalities? This isn't just a theoretical; it's already happening, says Sandra Matz, an associate professor at Columbia University. And she joined the Visiting Expert podcast to talk about it and her new book Mindmasters: The Data-Driven Science of Predicting and Changing Human Behavior. 

Three takeaways include

1. What you do online says so much more about you than you realize. Even the words you use in social posts can be telling about your personality. If you use lots of self-centred language (I, me, myself), it can actually be a sign of emotional distress. And studies from nearly a decade ago show that by simply analysing a person's Facebook likes and follows, an AI can be better at predicting someone's personality than their closest friends and family. 
2. And businesses can use this information to influence their consumer behaviour. Matz worked on one study with a beauty retailer that proves the point. For outgoing extroverts, language on the ads highlighted the fun nights ahead, while for introverts the ads focused on self-care at home, and the campaign was hugely successful. But Matz warns, leaders need to be very cautious about using this type of information; after all trust is easily broken and it's a fine line between giving people exactly what they want and need from a company and manipulation. 
3. Finally, when it comes to privacy issues, Matz warns it's not your social posts people should be most concerned with. It's our phones. The apps on our phones often have complete access to our photo galleries, location data, and even message history. Managing your data privacy can be a full-time job, but for folks who want to take a first step, she recommends being way more mindful about the permissions you give to companies when downloading your next app or game. 

New episodes of Rotman Visiting Experts are released monthly on Spotify, Apple Podcasts, and Amazon. Desktop listeners can also tune in on Simplecast.

Prefer to read than listen? Access the full transcript on Simplecast. 

If you enjoyed this podcast and aren't already a subscriber consider signing up for our bi-weekly newsletter, which delivers fresh thought-leadership and research straight to your inbox every other week.