You are here

What Transparency Reports Can Tell Us about Platform Engagement with Democratic and Autocratic Regimes

The final speaker in this Social Media & Society 2024 session is Sergei Pashakhin, whose is in the interfacing between platform companies and political institutions, especially in the context of autocratising regimes. Social media platforms operate around the world, and have to respond to the political and legislative situations in the countries in which they operate. Their transparency reports tend to provide a window into how they do so: these reports cover state requests for content moderation and take-downs, for instance.

Such transparency reports are driven in part by US lawmakers’ concerns about platforms’ arrangements with autocratic regimes (such as China), with respect to state interference in freedom of speech, surveillance, privacy, and other aspects. These reports can be seen both as addressing human rights standards and reporting on state interference, and serve as one element within the threefold set of options available to platforms engaging with autocratic regimes: loyalty to the regime, exit from the country, or voice about the state’s actions.

The overall assumption here is that platforms will receive more compliance requests from autocratising states or outright autocracies than democracies or democratising states, and will comply less with such requests than they do in democracies. The study tested this by reviewing transparency reports from Google, Meta, and Twitter from 2013 to 2022, finding that autocratising countries make significantly more requests than democracies; and that compliance here is lower than in democracies, too – but overall, compliance anywhere tends to be below 50%. (There is a complicated distinction between autocracies and autocratising countries here, too, which isn’t very clear to me.)

However, transparency reports remain a very limited source of information about platform companies’ engagement with governments, and should be taken with a grain of salt.