Social media platforms including Meta’s Facebook and Instagram, Twitter, TikTok and Google’s YouTube are readying themselves for another heated Election Day this week.

The companies now regularly come under close scrutiny around election time, something that accelerated following findings that Russian agents used social media to sow division in the run-up to the 2016 election. During the last presidential election in 2020, the platforms faced the challenge of moderating election denialism as an outgoing president stoked the false claims himself, leading several of them to at least temporarily suspend him after the Jan. 6 insurrection.

This year, the platforms are using all of those experiences to prepare for threats to democracy and safety as voters decide who will represent them in Congress, governor’s offices and state legislatures.

Here’s how all the major platforms are planning to police their services on Election Day.

Meta’s Facebook has been one of the most scrutinized platforms when it comes to misinformation. In response to years of criticism, it has bolstered its approach to election integrity. It’s said it will use many of the same policies and safeguards this year that it had in 2020.

Meta has stood up its Elections Operations Center, which it likened to a command center, to bring together different teams throughout the company to monitor and quickly address threats they see on the platform. It’s used this model dozens of times worldwide since 2018.

Facebook and Instagram also share reliable information with users about how to vote (including in languages other than English). The company said it’s already sent more than 80 million election notifications this year on the two platforms.

The company uses third-party fact-checkers to help label false posts so they can be demoted in the algorithm before they go viral. Meta said it’s investing an additional $5 million in fact-checking and media literacy efforts before Election Day.

Meta said it’s prepared to seek out threats and coordinated harassment against election officials and poll workers, who were the subject of misinformation campaigns and threats during the last election.

The company is once again banning new political ads in the week before the election, as it did in 2020. While ads submitted before the blackout period can still run, political advertisers have expressed frustration about the policy since it’s often helpful to respond to last-minute attacks and polling with fresh messaging. Facebook already has extra screening for those who sign up as political advertisers and maintains information about political ads in a database available to the public.

Meta has pledged to remove posts that seek to suppress voting, like misinformation about how and when to vote. It also said it would reject ads that discourage voting or question the legitimacy of the upcoming election.

In a study by New York University’s Cybersecurity for Democracy and international NGO Global Witness testing election integrity ad screens across social media platforms, the groups found Facebook was mostly successful in blocking ads they submitted with election disinformation. Still, 20% to 50% of the ads tested were approved, depending on what language they were in and whether they were submitted from inside or outside the U.S.

The researchers also violated Facebook’s policies about who is allowed to place ads, with one of the test accounts placing ads from the U.K. The researchers also did not go through Facebook’s authorization process, which is supposed to provide extra scrutiny for political advertisers.

The researchers did not run the ads once they were approved, so it’s not clear whether Facebook would have blocked them during that step.

A Meta spokesperson said in a statement published with the study that it was “based on a very small sample of ads, and are not representative given the number of political ads we review daily across the world.”

“We invest significant resources to protect elections, from our industry-leading transparency efforts to our enforcement of strict protocols on ads about social issues, elections, or politics – and we will continue to do so,” a Meta spokesperson said in a statement to CNBC. The New York Times first reported the statement.

“TikTok is first and foremost an entertainment platform,” the company said in a September blog post. It added that it wants to “foster and promote a positive environment that brings people together, not divide them.”

Still, the NYU and Global Witness study found TikTok performed the worst out of the platforms it tested in blocking election-related misinformation in ads. Only one ad it submitted in both English and Spanish falsely claiming Covid vaccines were required to vote was rejected, while ads promoting the wrong date for the election or encouraging voters to vote twice were approved.

TikTok did not provide a comment on the report but told the researchers in a statement that it values “feedback from NGOs, academics, and other experts which helps us continually strengthen our processes and policies.”

The service said that while it doesn’t “proactively encourage politicians or political parties to join TikTok,” it welcomes them to do so. The company announced in September that it would try out mandatory verification for government, politician and political party accounts in the U.S. through the midterms and disable those types of accounts from running ads.

TikTok said it would allow those accounts to run ads in limited circumstances, like public health and safety campaigns, but that they’d have to work with a TikTok representative to do so.

TikTok also barred these accounts from other ways to make money on the platform, like through tipping and e-commerce. Politician and political party accounts are also not allowed to solicit campaign donations on their pages.

TikTok has said it’s committed to stemming the spread of misinformation, including by working with experts to strengthen its policies and outside fact-checkers to verify election-related posts.

It’s also sought to build on its experiences from the last election, like by surfacing its election center with information about how to vote earlier in the cycle. It’s also tried to do more to educate creators on the platform about what kinds of paid partnerships are and are not allowed and how to disclose them.

Twitter is in a unique position this Election Day, after billionaire Elon Musk bought the platform and took it private less than a couple weeks before voters headed to the polls.

Musk has expressed a desire to loosen Twitter’s content moderation policies. He’s said decisions on whether to reinstate banned users, a group that includes former President Donald Trump, would take a few weeks at least.

But shortly after the deal, Bloomberg reported the team responsible for content moderation lost access to some of their tools. Twitter’s head of safety and integrity, Yoel Roth, characterized that move as a normal measure for a recently acquired company to take and said Twitter’s rules were still being enforced at scale.

But the timing shortly before the election is particularly stark. Musk said teams would have access to all the necessary tools by the end of the week before the election, according to a civil society group leader who was on a call with Musk earlier in the week.

Before Musk’s takeover, Twitter laid out its election integrity plans in an August blog post. Those included activating its civic integrity policy, which allows it to label and demote misleading information about the election, sharing “prebunks,” or proactively debunked false claims about the election, and surfacing relevant news and voting information in a dedicated tab. Twitter has not allowed political ads since 2019.

Like other platforms, Google and YouTube highlight authoritative sources and information on the election high up in related searches. The company said it would remove content violating its policies by misleading about the voting process or encouraging interference with the democratic process.

YouTube also has sought to help users learn how to spot manipulative messages on their own using education content.

Google said it’s helped train campaign and election officials on security practices.

Clarification: This story has been updated to clarify that Meta’s statement on the study was first reported by The New York Times.

Subscribe to CNBC on YouTube.

WATCH: The messy business of content moderation on Facebook, Twitter, YouTube