Ever since Russian agents and other opportunists abused its platform in an attempt to manipulate the 2016 U.S. presidential election, Facebook has insisted — repeatedly — that it’s learned its lesson and is no longer a conduit for misinformation, voter suppression and election disruption.

But it has been a long and halting journey for the social network. Critical outsiders, as well as some of Facebook’s own employees, say the company’s efforts to revise its rules and tighten its safeguards remain wholly insufficient to the task, despite it having spent billions on the project. As for why, they point to the company’s persistent unwillingness to act decisively over much of that time.

The company’s rhetoric has certainly gotten an update. CEO Mark Zuckerberg now casually references possible outcomes that were unimaginable in 2016 — among them, possible civil unrest and potentially a disputed election that Facebook could easily make even worse — as challenges the platform now faces.

“This election is not going to be business as usual,” Zuckerberg wrote in a September Facebook post in which he outlined Facebook’s efforts to encourage voting and remove misinformation from its service. “We all have a responsibility to protect our democracy.”

In the immediate aftermath of President Donald Trump’s election, Zuckerberg offered a remarkably tone-deaf quip regarding the notion that “fake news” spread on Facebook could have influenced the 2016 election, calling it “a pretty crazy idea.” A week later, he walked back the comment.

Since then, Facebook has issued a stream of mea culpas for its slowness to act against threats to the 2016 election and promised to do better. “I don’t think they have become better at listening,” said David Kirkpatrick, author of a book on Facebook’s rise. “What’s changed is more people have been telling them they need to do something.”

The company has hired outside fact-checkers, added restrictions — then more restrictions — on political advertisements and taken down thousands of accounts, pages and groups it found to be engaging in “coordinated inauthentic behavior.” That’s Facebook’s term for fake accounts and groups that maliciously target political discourse in countries ranging from Albania to Zimbabwe.

It’s also started added warning labels to posts that contain misinformation about voting and has, at times, taken steps to limit the circulation of misleading posts.

All this unquestionably puts Facebook in a better position than it was in four years ago. But that doesn’t mean it’s fully prepared.

Copyright 2020 The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed without permission.