Where tech aligns

Sliding into Pavlov’s DMs

Learning models can help explain why people are replacing face-to-face interactions with screens.

Even without any notifications nudging you, you are almost certain to be assaulted by an itch to check your phone throughout your day. Decades of psychology research that began with measuring dog saliva have produced models that explain how that itch has developed. It is the convenience of social media that influences our psyches more than anything else. We all have social needs, and we’re increasingly fulfilling those with the social equivalent of empty calories. 

Our motivational biologies are similar enough to animals used in academic research that we can predict what brain regions will be active in tasks in humans from looking at animals. And we can even do the converse: we can predict what brain regions animals will use in tasks originally developed in humans. This is because human and other animal minds have a shared brain blueprint that evolved to explore and exploit their environments. We’ve evolved to be miserly, lazy, and hedonically shortsighted. We all expend as little energy as possible to maximize return on energetic investment. A smart animal asks: What in my environment provides the easiest, best rewards?

For us humans, other people tend to win the prize for being most easily associated with powerful rewards. We’re social animals. This is because individually we’re weak. To have been alone in the wilderness of our evolutionary history – a history instantiated in the DNA that forms our brains and minds – meant death. When we were alone, we stood little chance to feed ourselves or fend off predators. But as a group, we can do so much more.

The company of others became a signifier for safety, food, shelter, and comfort. And as a result, others take on a reward valence as an end in themselves. We crave interaction with others, similarly to how we crave food and play. But the neural circuitry that is responsible for these cravings has been tapped into by technologies that befuddle us, essentially tricking us into believing that the little glowing screens we surround ourselves with are providing the same kind of boon that the company of others does.

The picture, as it pertains to how digital tech has hijacked our social monkey-minds, is this: It feels good to spend time with people face-to-face. It’s uncontroversial way more rewarding to be with people in person. And yet, despite that, people have been spending less time in person. (That’s even according to data from before the pandemic.)

It’s obvious that we feel much more connected when together in-person; ask anyone who tried to use a video call during the Covid lockdowns for some ersatz happy hour. It sucks. There’s something missing, and it’s palpable. But people still did these because in spite of the insufficiency, half-presence with another is preferable to being alone.

This is because the suffusion of tiny digital social rewards has led to an unconscious replacement of actual meatworld interaction with social media use. To be able to pick up your phone and see umpteen different photos of familiar faces, funny or insightful posts that you relate to, and see others responding to your own content is a nice little social interaction. And it’s always there. Some days it might be less rewarding than others, but that actually increases the appeal

Despite what we think about ourselves, we’re really not that smart. Our monkey-minds don’t know that the notifications, “Likes,” and comments are just shoddy replacements for the sort of stimuli that are ethologically relevant and actually meaningful. Our monkey-minds didn’t evolve to discern whether an interaction really took place in person or on a smartphone. Data show that we feel the very real pain of social exclusion even in virtual social contexts. As I said, we’ve evolved to be lazy animals. So if we’re able to get some lower level of reward, or to avoid some perception of being left out, we eventually learn that this is a good, easy way to get SOME kind of stimulation that feels nice because it’s some kind of social interaction. Again, this is happening because our brains were not evolved to discern real, in-person interactions from those that take place in cyberspace.

There is a model, the Rescorla-Wagner model, that does a fair job at describing how this kind of second-order conditioning can work:

ΔVαβ ( λ – ∑V )

It was developed to describe how changes in associative strength (ΔV) are influenced by the salience of environmental stimuli (α) and the valence of the outcome they predict (β). In this case, the environmental stimuli are whatever social information is appearing on your phone, and the valence there is whether that feels good or bad and just how good or bad it feels. And in broad strokes, it feels good to just have frictionless social interaction (like scrolling through or putting your ideas onto a platform to receive “Likes”).

This and other similar models were developed in carefully controlled lab settings to ask the following question: How do animals (and by extension, humans) learn about the specific elements of their environments that lead to either rewards or bad outcomes? In lab settings, these “specific elements of their environments” are things like levers and noises that are paired with either an aversive outcome (like a mild footshock) or a reward (like a food pellet or even access to interact with another lab rat). Phones are about as natural to humans as levers are to lab rats, so it’s not a stretch to say we can use models developed in the latter to understand the former. Again, there’s a mountain of empirical validation that the same principles found in lab animals apply to humans.In essence, I’m using the Rescorla-Wagner to say that strong and weak rewards can attain the same levels of associative learning as one another – just the rate differs.

So with enough exposure you’re bound to form a pretty strong association of phone = socialization.

For the sake of my argument, I’m assuming that the stimuli (α) are the same and just the reward strength (β) is different.

In the graph above, the Y-axis is the strength of the association, and the X-axis is time or exposure to the stimuli. Ultimately, with enough exposure, they elicit the same level of association between the stimulus and the reward. So, by dint of repeated exposures to the rewarding elements of little human-like images in your phone, there may eventually come a time when you behave as if looking at a mobile phone is a replacement for spending time with a pal or your grandmother. Your monkey-mind is miserly. It’s easier to pick up a phone than to drive a town away to visit an old friend or to bike fifteen blocks to meet up at the bar, so that’s what people have been doing. And this calculation is happening without conscious awareness – people don’t even realize that they’re making a choice. 

That neural circuitry intended for socializing is being hijacked by parasocializing media is nothing new. The ever-glowing television set has acted as a cornerstone of isolated comfort for years, and even reading a book will make you feel less lonesome. But there’s something distinct about interfacing on social media that seems to be particularly convincing to us. It’s seductive in a way other media are not. And centrally, the subjective value of, and motivation for, rewards is modulated by need states. If you’re hungry, the smell of eggs and bacon sings a song sweeter than any other. And if you’re lonely, even the echoes of friendship feel better than nothing else. In the case of social media, the cycle is self-perpetuating: in lieu of actual interaction, it’s easier to just SCROLL.

Pulling back the wiring on how we lose agency won’t automatically rescue us from the clutches of bad habits. But it’s a start.

Lots more where that came from.