Aaron Sorkin’s new movie The Social Reckoning, the hotly anticipated follow-up to The Social Network, tells the story of whistleblower Frances Haugen, a Facebook employee who showed that the company knew its platforms harmed some users.

Haugen, played in the film by Mikey Madison, shared Facebook’s internal data with journalist Jeff Horwitz (played by Jeremy Allen White) at the Wall Street Journal. The resulting investigation, “The Facebook Files,” was published in 2021. It and other reporting led to congressional investigations, lawsuits and more action against Facebook—much of which is ongoing.

It also gave urgency to research investigating how social media affects people’s mental health, especially that of children. Among these scientists is Laura Edelson, a social media safety researcher at Northeastern University, who has studied Facebook’s moderation practices during the January 6 insurrection, the way that misinformation spreads on Facebook and social media privacy across multiple platforms. We took Edelson to see The Social Reckoning to find out what the film gets right about social media’s potential harms.

On supporting science journalism

If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

“It was really, frankly, surreal seeing a series of events that you lived through on the big screen,” she says.

[An edited transcript of the interview follows.]

In the movie, we learn about Facebook’s 2018 transition to an algorithm based on “meaningful social interaction” (MSI), which had disastrous consequences for some users. What is MSI, and how does it work?

Before 2018, Facebook primarily maximized [its algorithm] for duration of use. They cared about how long your Facebook session was and how many times [you returned] over a day or a week. MSI [maximizes for] users who are interacting, who are clicking things, who are typing. It is very different from a passive, just viewing, just scrolling state.

When you are clicking a lot of things and you’re responding a lot, you can see how that could apply both to congratulating someone on their new puppy or [to commenting], “This is the dumbest idea ever. You should go die.” As much as those two [actions] are very different things, you’re actually in a similar state of being deeply engaged and paying a lot of attention. That is obviously very, very different than an algorithm that just tries to get you to be there for as long as possible. It’s going to select different things to show you.

Content that is going to be very interactive is going to be the kind of content that is more extreme content. And that’s true whether it’s something that makes you happy or whether it’s something that makes you angry or scared.

Why did Facebook make the shift to MSI?

The why is because that is what they sell. Facebook sells ads. But more specifically, what they sell is user interaction with ads. That’s not to say that no Facebook ads are sold on a per-view basis, but mostly it’s this interactivity, this click-through rate, that’s the thing that is important to Facebook’s bottom line.

The movie seems to cast blame on Facebook for increasing political radicalization in 2019–2021, including the January 6 riots. Do you think it fairly portrays Facebook’s role?

New mediums have profoundly destabilizing effects on political structures. I think it is not surprising that social media had a similar effect. It’s hard to wrap your head around because what happened was absolutely an amplification of existing underlying patterns. But there is a point at which moderate content literally cannot compete—that becomes a point at which, if you are a politician, it is genuinely difficult to spread a moderate political message because it will perform badly on the dominant medium of communication. And that’s not a left thing or a right thing. That [means that] the more extreme your message is, the better it will perform on both social media. I see that repeatedly in every study I ever do of content.

In one study we observed that on January 6, a lot of [content posted to Facebook] got taken down. And then, on January 12, a larger tranche of stuff was taken down. The thing was: the vast majority of these takedowns happened too late to have much of an effect [on what people saw in their news feeds]. It had some very small impact, but moderating after the content has already been up for days just doesn’t change user experience on your platform very much.

It matters not just what you moderate but when you moderate. If your fact-checking takes several days, if your system that identifies content that might violate your rules takes several days, you might as well not do it just because it’s not going to matter much in terms of user experience.

Several studies suggest that social media shows people more extreme content but does not necessarily change their views. Is that true?

It’s very hard to set up an experiment that is causal, and that is what you’d have to do to answer that question. You need to set up a controlled experiment where you take a population and then you randomize and then you show them different stuff. An experiment from this year did that in a long-term way. And the researchers actually did find an effect. They found that when [the researchers] changed what was in someone’s feed, [the participants’] political beliefs changed.

This movie depicts Facebook in 2021. How have social media algorithms changed since then?

Things are quite different in 2026. That is not to say that some of the problems that were very visible in 2021 are not still with us, but one of the things that has happened is: Facebook is a much smaller and more marginal place. TikTok, Instagram Reels and YouTube Shorts are the places where more people are spending more of their attention.

[Political content] is not as uniquely advantaged [on those platforms] as it was on Facebook [because of] the difference in monetization of how ads are sold on short-form video platforms. They are more likely to be sold on a per-view basis. The algorithm for TikTok is much more durational. To keep someone on a platform of that type for as long as possible, it’s going to wind up favoring different kinds of things.

It’s creating different kinds of problems; [for example], the eating disorder problem continues to be very real. There remains a big conspiracy theory problem, particularly around things that get monetized. But the political extremism problem is not as severe—not because anybody really did anything [to address] it but because this other form of social media that just was able to capture people’s attention longer and better became more dominant, and it’s prone to different problems.

What are your final thoughts on the film?

It got a lot of small things that I knew about right. The movie made it very legible why Jeff did what he did and why Frances did what she did. But I did not find that movie very enlightening as to why Mark Zuckerberg does what he does.

I think we all like to think, when there are such obvious bad things happening in the world, that it’s because there’s a villain twirling their mustache and making the bad thing happen.

But I think a lot of bad things in the world happen because good people let them. Or good people have other reasons for finding that bad thing to be tolerable. I think it would be more useful to understand that as a way of getting to “How can we make social media safer for people?”