Twitter admits there were many more Russian trolls on its site during the 2016 U.S. presidential election

Twitter revealed on Friday that trolls tied to the Russian government spread far more disinformation during the 2016 U.S. presidential election than the company first reported — and it pledged to notify hundreds of thousands of users who had seen that content.

The update comes as Twitter continues to face criticism on Capitol Hill that it has failed to fully confront the scourge of Kremlin propaganda — and neglected to respond to the earlier demands of lawmakers who are probing Russia’s meddling on popular social media sites.

Ahead of a series of congressional hearings last year, Twitter initially said it had discovered 2,200 accounts tied to the Internet Research Agency, a troll army connected to the Russian government. On Friday, though, Twitter said it had actually identified 3,814 accounts related to the IRA.

Also last year, Twitter calculated that there were roughly 36,000 bots originating out of Russia — and tweeting about the election — as Americans headed to the ballot box. By Friday, though, Twitter said it had found an additional 13,000 bots, bringing the total tally of automated accounts tweeting about the presidential race to more than 50,000.

And Twitter revealed for the first time on Friday that Russian propaganda — content that sought to stir social and political unrest in the United States — reached scores of its users. The company said it would notify 677,000 people in the United States who had followed one of these suspect accounts, or retweeted or liked their content. Twitter said it would do so by email.

In announcing its findings, Twitter sought to stress that Russian disinformation only amounted to a small portion of the tweets shared regularly on its platform. And it reiterated that it had taken steps to prevent such abuse as another election — a 2018 race to determine the composition of Congress — fast approaches. That includes a series of previously announced changes to the way it displays political ads.

But the news is sure to infuriate some federal lawmakers, who repeatedly have needled Twitter during the course of their investigation into Russian influence.

Democratic Sen. Mark Warner, for one, blasted Twitter in September for a “deeply disappointing” response to his questions about the election. When the company later appeared with its tech peers, Facebook and Google, at a series of congressional hearings on the issue, lawmakers from both parties demanded that Twitter take more aggressive steps to prevent such manipulation of its platform in the future.

This year, the company completely blew a deadline by which it was supposed to respond to written questions it was sent by congressional investigators. And for months, Twitter had ignored public demands by lawmakers like Sen. Richard Blumenthal that it notify users who had seen or interacted with such Russian propaganda. Only this week did Twitter finally acknowledge that it would take that step.

Warner, for his part, still offered some limited praise late Friday:

Blumenthal, meanwhile, cheered Twitter’s belated decision to notify users who had seen content generated by Russian trolls. But the Democratic senator still said he’d keep watch to ensure that Twitter actually adopts “measures to implement safeguards to protect users from the ongoing and real-time influence of Russian bots.”

Nevertheless, Twitter announced its latest findings at a busy moment: The company published its blog post at 5 pm on a Friday, while the U.S. Congress barreled at the time toward a potential government shutdown.

This article originally appeared on Recode.net.