Will New EU Law Begin Holding Social Media To Account?

More than a year and a half after the storming of the Capitol on January 6, the misinformation that fueled the riot continues to proliferate on mainstream social media. In the fall-out, sites like Facebook and Twitter took steps to remove false content and ban accounts that violate their terms of service, but enforcement policies remain ineffective, opaque, and erratic. YouTube, for example, took down a clip posted by the House Committee investigating January 6, arguing that the committee’s video itself was spreading misinformation about the election.

Despite the longstanding prevalence of the misinformation problem online, social media companies have yet to enact anything approaching effective measures to curb it. Self-regulation clearly isn’t working, but political gridlock and polarization have thus far dampened prospects for government intervention in the US.

The European Union, however, recently took a significant step in passing the Digital Services Act (DSA), which policymakers approved in April. (The law will take effect next year after a final vote, but significant changes are not expected.) While the law isn’t perfect, policymakers in the US — and around the world — should be paying close attention.

Let me outline some important features of the new legislation and potential consequences. First, the law tightens content moderation requirements. Large-scale social media platforms will be required to develop concrete plans to remove illegal content from their platforms. They will also be subjected to stricter reporting requirements, closer cooperation with regulators, and new risk management and response guidelines.

These provisions could bring some order to a content moderation system currently in disarray, but they are not the best feature of the new law.

From my perspective, the most promising aspect of the law are the new rules that govern platforms’ interactions with users. Significantly, platforms will have to publicize details about the algorithms that recommend content to users, as well as giving them the option to opt out of these recommendation systems entirely.

This new transparency around algorithms should be celebrated. There is good evidence that these algorithms can destroy attention and mental health; create dependency and addiction; and accelerate extremism. At the very least, the public must be shown how these algorithms work and given the option to use these platforms without being subjected to purposefully addictive feeds.

Social media networks will also be prohibited from using sensitive data — such as user’s race, ethnicity, religion or beliefs, disability, or sexual orientation — to target them with advertisements, nor will they be allowed to serve targeted ads to kids.

Finally, the law imposes requirements that platforms share more data with researchers so they can assess risks and trends online rigorously and independently. This is an especially important provision given the alarming results of proprietary, internal research carried out on platforms like Facebook. An internal Facebook analysis of teenage girls’ use of Instagram, for example, found that the platform is toxic for mental health and body image. There is no reason the public — which, after all, generates all this data — should rely on internal leaks and whistleblowers for a full accounting of these platforms’ negative effects.

Not everyone likes the EU’s Digital Services Act. Critics say there is also a risk that the new content moderation regulations are too rigid and could backfire, leading to even more erratic moderation.

Even more seriously, there are legitimate questions over whether the law can be adequately enforced. A 2016 internet privacy law, the General Data Protection Regulation, was heralded when it was passed, but so far has had a limited impact in practice, in part because of inadequate and slow enforcement mechanisms. While the DSA will be enforced by the European Commission itself, its staff for enforcing these laws pales in comparison to the resources of the social media giants it is trying to regulate.

In aggregate, the DSA represents an important step forward. Content moderation is a very thorny problem, and it’s unlikely that the DSA will represent the final word on solving it. Policymakers may have to use their imaginations to explore solutions in a domain that is constantly evolving and expanding. Still, at the very least, this attempt will prove instructive and help policymakers learn important lessons and develop new ideas.

To me, the most serious harms from social media come from the way they can rob people of agency. Hopefully, the law will result in new insights into how platforms are affecting people’s judgment and mental health. I even hold out some measure of hope that, down the line, the DSA and similar efforts could lead to some changes to social media companies’ predatory, “addiction-by-design” business model, and help us use social media more mindfully and productively.