Ann M. Ravel is the Digital Deception Project Director at MapLight and previously served as chair of the Federal Election Commission. The opinions expressed in this commentary are her own.

As the 2020 election nears, the big three social media companies have been scrambling to solidify their approaches to political advertising.

Last month, Google, which owns YouTube, announced that it would limit certain types of political ad targeting, but made no change in allowing politicians to run false ads. Twitter made waves by announcing that it would ban certain types of political advertising on its platform. And Facebook announced that it would not fact-check ads by politicians — but would still fact-check other advertisements.

While Twitter’s gamble seems to be paying off in the court of public opinion, Google’s and Facebook’s are receiving criticism. In the long run, however, all of the platforms’ decisions are bad for democracy and show why social media companies shouldn’t be self-policing when it comes to political ads.

On the surface, Twitter’s announcement may seem like a proactive way to deal with a broken digital political advertising system. However, there are many problems that indicate that Twitter should reverse its ban on political candidates, PACs and SuperPACs. It is likely, for example, that the ban will favor incumbents and establishment candidates with deeper pockets (who can afford television advertising) over smaller challengers.

Since digital ads are cheaper than radio or television spots, non-incumbents often rely on them to reach voters, mobilize volunteers and raise money. Moreover, Twitter still plans to allow certain groups to advertise on what they consider to be political issues — but only so long as they do not advocate for or against political or legislative outcomes. This is likely to create enforcement challenges and cause confusion for groups seeking to educate the public about issues.