February 19, 2020 - 3:55pm

What do they know about us? The title of a new BBC documentary, about Amazon, gives away a key concern of our age. What do the big tech companies — Amazon, but also Facebook, Apple, Google, Twitter, Netflix — know about us? What do they share, what do they keep, how well can they predict us and track us?

I can be blasé about it sometimes — my life is so vanilla that I don’t care who knows what I get up to, and I do my “how to murder, best murder methods, disposing of bodies” searches in an incognito window via DuckDuckGo anyway. But it is a concern.

There is another concern, though. And that’s not that they have too much data on us, but that we have too little on them.

As well as the privacy concerns, we’re constantly worrying about whether or not Twitter makes us angry or Instagram makes us sad or using screens stops us sleeping well; whether we’re addicted to our phones; whether social media spreads racist views or political division, or is tied to poor mental wellbeing; whether video games cause violence.

I have written in the past about why those claims are probably overblown. The studies are weak, they rely on people’s memories of how much they use screens which hardly line up with real use, and they’re usually p-hacked to pieces. Big, careful studies tend to find small, uncertain correlations; there’s not much evidence that tech is harming us.

But even those big, careful studies mainly rely on self-reported data; some now use apps that record phone use, but that’s only starting. And more than that, it’s expensive to gather. Any subtle health impacts might be hard to find, because slight correlations won’t show up without huge datasets.

Big tech companies, though, have those huge datasets. They have the info on when people came into their apps or sites, when they left them, how they used the phone and how many times they opened it, and so on. It’s not that it would immediately make all the research easy — you’d still need to gather data on mental wellbeing, or whatever your dependent variable is — but it would be hugely beneficial.

As I understand it, there have been talks about Google and Facebook sharing information, but progress is slow. At the moment, most studies about the political impact of social media are on Twitter, because — unlike Facebook — it’s searchable. It’s the drunk looking for his keys under the lamppost, not because they’re there but because that’s where the light is.

Tech firms’ whole business model is based on our data, so obviously they’d be reluctant to share all of it. But presumably they could share some without major cost. And since they profit off our data, sharing it with researchers in order to help us seems the least they can do. What they know about us is one thing; I’m more interested in what we know about them.


Tom Chivers is a science writer. His second book, How to Read Numbers, is out now.

TomChivers