The first step of taking on social media companies is to develop much more comprehensive rules that platforms must conform to. Ideally, that process would be done collaboratively with likeminded countries cooperating on a set of enforceable standards that apply across markets like Europe, Canada, Australia, Japan, Brazil, and beyond to make it harder for tech companies to wield their leverage against an individual country going it alone. Those rules should be comprehensive, looking at all aspects of platform design and operation. They could include reining in dark patterns that manipulate users, limiting if not outlawing aspects of algorithmic amplification of content, implementing high standards of moderation and larger penalties for breaches, and enforcement mechanisms that allow swift action when rules are violated, leaving the door open to blocking platforms that do not comply — as Brazil did to X in 2024.

The goal of these standards would be to limit the harmful effects of social media platforms, but that will naturally force them to change how they look and operate to conform with those restrictions. Social media companies have long argued that they are platforms, not publishers, to try to avoid accountability and their obligations to their users’ wellbeing. But when they develop complex algorithmic systems that elevate particular content in front of the eyes of their users — including, but not limited to, news content — and even incentivize the creation of certain kinds of content, that is clearly not the case. The standards they abide by must be raised, regardless of the protests that will come from tech billionaires and their allies in the US government.

Reining them in is just the first step. Paired with that must be a serious, well-financed effort to develop alternative means of digital communication that rethink the social media model from the ground up, built on entirely different foundations than the private, profit-hungry platforms that now shape online discourse. This should be approached from many different angles, with the recognition that some of the projects may fail, but that’s acceptable because experimentation in a short timeframe is necessary.