Fact-checkers are out. The internet gets to vote on the truth now.

By Ren LaForme, managing editor

Once upon a time — back in the hazy days of 2016 — platforms like Facebook and Twitter looked inward and spotted a problem. Their systems were infected with a wicked, fast-spreading virus. It was time to do something about it.

The problem? Mis- and disinformation. Election hoaxes. Distortions of basic science. Wild claims like the one that a pizza place in Washington, D.C., was secretly hosting a child sex trafficking ring.

The solution: Bring in professional fact-checkers — trained journalists who know how to separate fact from fiction.

In the systems that followed, social platforms used fact-checkers’ work to flag misleading posts, label them with context and reduce their reach. The fact-checkers never had access to the platforms themselves, only to the claims.

It wasn’t perfect, but it worked — or at least, it helped. Then, in January, the backlash reached a fever pitch.

Some folks in D.C. cried censorship. They pressured platforms to sever ties with fact-checkers. And in many cases, that’s exactly what happened. Twitter/X and Meta both moved on from those relationships, replacing them with crowdsourced systems like Community Notes — which, promising as they sound, have more than a few flaws.

This week, TikTok became the latest platform to follow suit, announcing it would launch a Community Notes-style system alongside its existing fact-checking partnerships. So I caught up with MediaWise director Alex Mahadevan, Poynter’s resident expert on community notes, to ask: How’s this all going?

The following conversation has been edited for brevity and clarity.

Ren LaForme: Why are so many platforms moving away from professional fact-checkers and toward community notes?

Alex Mahadevan: X owner Elon Musk has been hostile toward journalists for a while. And in January, Meta’s Mark Zuckerberg trashed fact-checkers with misleading statements about bias and mistakes they’ve made. So it’s partially politics — Musk’s sincerely and Zuckerberg’s cynically, in my opinion. They’re handing the mic to you, the user, so you don’t have to rely on those pesky reporters anymore.

Relatedly, by handing fact-checking over to users, platforms dodge accountability and insulate themselves from claims of bias or censorship. Again, politics.

But it’s mainly about the bottom line, if you ask me. Fact-checking costs money. And these crowdsourcing systems are much cheaper to build and maintain. Users do it for free! They can throw out all kinds of justifications, but in the end, it’s about money.

LaForme: Concerning! What does this say about how social media companies view their role in moderating content?

Mahadevan: They are saying that this is about freedom of speech. And in its announcement about ending fact-checking in January, Meta specifically mentioned the need to relax most moderation. It says that by giving the power to assess truth to their users, the platform proves its commitment not to be “the arbiters of truth.”

That’s a total mirage.

Meta is still building the systems that power community notes. The company chooses whether to put a 0.5 or 0.6 in the algorithm that determines whether a note actually appears on a post. They claim this bridging algorithm — the consensus model that surfaces a community fact check if it’s upvoted by a wide variety of perspectives — is the best way to determine truth. But why? Again, the company is what decided it.

I still think facts are facts. Reality is reality. It’s not a partisan issue. And the platforms seem determined to make it one.

LaForme: Right. They’re offloading the perception that they have a finger on the scale while keeping all five firmly planted there. So what do these platforms gain by outsourcing fact-checking to users?

Mahadevan: Aside from ducking responsibility and cashing in on hefty savings, X and Meta say speed and scale. Community members can quickly flag and add context to misleading posts. And more contributors mean more fact checks, right?

But the numbers don’t really add up. Instagram’s Adam Mosseri said last month that U.S. fact-checkers managed 100 checks per day. He didn’t mention that those fact checks were appended to thousands and thousands of related posts on Facebook and Instagram.

On X, contributors manage roughly 250 public community notes per day — a number that has taken a sharp dive since X tweaked the algorithm to make it more difficult for notes to appear.

As for speed, research has found that some community notes do indeed appear quickly, but not quickly enough to curb the reach of harmful misinformation.

LaForme: Crowdsourcing, in my experience, is a wild ride. You can get some really great, unexpected stuff, but you’re often dealing with piles and piles of chaff. Are there examples of community notes going very right — or very wrong?

Mahadevan: I am glad you asked this, because I don’t want to turn this into a community notes roast. I love the idea of crowdsourced fact-checking. And research has shown that people trust community fact checks more than the pros.

On X, Community Notes contributors are good at quickly flagging misleading advertisements and AI-generated slop. There is lots of agreement on those items, so notes appear quickly — faster than a professional fact-checker could manage, I’ll admit. I have seen lots of fake disaster footage debunked quickly.

But, when it comes to harmful misinformation, community notes fall short. The most prominent sharers of misinformation on X are also the most polarizing. That means it’s nearly impossible to get a community fact check appended to a post from someone like Musk.

Back to the positives: X’s Community Notes is radically transparent. The algorithm and data on every note and rating are available to download. That’s what drew me to the system in the first place. I was sitting in a coffee shop, messing around with Birdwatch data.

I’m always impressed with how quickly the team behind Community Notes responds to user feedback. They’ve pushed a lot of meaningful changes to the system over the last few years. So I hate ragging on them. To be clear, Community Notes is a brilliant system if it existed within a true trust and safety program.

And the bridging algorithm — the thing that requires consensus from all sides — would also be a great replacement for the newsfeed or “for you” tab. I'd love to see Meta and X put their money where their mouths are: If the bridging algorithms are so awesome, use them — instead of the polarizing, engagement-focused ones they have now.

LaForme: Could professional and crowdsourced fact-checking coexist in a meaningful way? Could TikTok's plan prove it? Or is that just lip service?

Mahadevan: Absolutely. That was how Birdwatch was supposed to work, according to Yoel Roth, former head of trust and safety for then-Twitter.

And if TikTok honors its commitment to keep professional fact-checkers engaged on the platform, Footnotes can prove to be the best model for crowdsourced fact-checking.

In the ideal world, fact-checkers can debunk the most harmful and viral misinformation, especially around complex political topics. Footnotes contributors can add context to everything else.

I don’t know whether professional fact-checkers will have any involvement in the Footnotes system, but ideally they could serve as a backstop when a note gets lots of upvotes but not enough consensus to appear on TikTok.

Also, on X and Meta, note quality and accuracy seem to be based on how many “helpful” votes they get. Professional fact-checkers could fill in the gap by regularly auditing notes to check source quality and accuracy.

LaForme: Do you think we’ll look back on this moment as a turning point in how we understand truth online?

Mahadevan: Yes. The tools to deceive people are cheaper and more powerful — think about generative AI slop — the rewards for deception are higher than ever — think about all that engagement farming on every platform — and the guardrails against online deception are much lower than they’ve been in years.

Platforms, with the rollout of community notes, are telling users: “Hey, it’s a hostile digital world out there, and you’re on your own.”

Part of me worries that trust in legitimate journalists and fact-checkers will continue to decline as people embrace community notes. But, then I remember: The highest-rated notes I see have links to the Associated Press, Reuters, PolitiFact, the BBC. Fact-checkers are among the top sources for community notes on X.

The platforms may think they’re leaving fact-checkers and newsrooms behind, but the users are telling a different story. The platforms need us.

Corso to retire from ‘GameDay’

By Tom Jones, senior media writer

Lee Corso, the former football player and coach who has been with ESPN’s “College GameDay” since its inception in 1987, is retiring from the show. ESPN announced Corso will do one final show — next season’s season debut on Aug 30.

Corso, who turns 90 in August, has scaled back his presence on the college football pregame show in recent years, but he has continued with the most famous part of the broadcast. “GameDay” concludes each show with a predictions segment, capped by Corso picking the winner of the game of the day by putting on that team’s mascot head. Because the show is live on a different campus each week, Corso’s much anticipated moment is either wildly cheered or booed by the crowd, depending on who he picks. According to ESPN, Corso has made 430 headgear picks all time. (By the way, he has picked Ohio State 45 times, more than any other school.) He has a pretty impressive record, too: 286-144.

Corso joined ESPN in 1987 after a 28-year coaching career, including college stops at Louisville, Indiana and Northern Illinois. He is the only original member still left on “College GameDay.”

Corso said in a statement, “ESPN has been exceptionally generous to me, especially these past few years. They accommodated me and supported me, as did my colleagues in the early days of ‘College GameDay.’ Special thanks to Kirk Herbstreit for his friendship and encouragement. And lest I forget, the fans … truly a blessing to share this with them. ESPN gave me this wonderful opportunity and provided me the support to ensure success. I am genuinely grateful.”

Fellow analyst Kirk Herbstreit said in a statement that Corso has been “like a second father to me.”

ESPN chairman Jimmy Pitaro called Corso, “one of the most influential and beloved figures in the history of college football.”

The Poynter 50

By Ren LaForme, managing editor

The seventh installment of The Poynter 50 published Thursday, continuing our yearlong series reflecting on 50 moments and people that shaped journalism over the past half-century — and still influence its future.

This latest piece, by Poynter media business analyst Rick Edmonds, revisits a moment in New York Times history that’s almost hard to imagine now: In 2009, facing the dual pressures of overexpansion and a deep recession, the Times accepted a high-interest loan from Mexican billionaire Carlos Slim.

It follows another Poynter 50 story published earlier this week — PolitiFact’s Louis Jacobson on Matt Drudge’s Monica Lewinsky scoop and the fall of traditional media gatekeepers.

Media tidbits and links for your weekend review

• An Associated Press photographer was allowed to join the White House pool Thursday, but the agency’s text reporters face continued exclusion. Both sides are scheduled to appear in court Friday to determine whether the White House is violating a court order requiring it to lift the ban on the AP. Meanwhile, a panel of three judges at the U.S. Court of Appeals for the District of Columbia Circuit is evaluating an appeal from the White House and heard arguments Thursday. The Hill’s Ella Lee has more in “Appeals court appears split on AP access to White House.”
• Rather than deterring President Donald Trump, critical journalism has seemingly encouraged him to double down. Vanity Fair’s Gabriel Sherman explores "How a ‘No-Scalps Policy’ Is Shaping Trump’s Consequence-Free Second Term.”
• “US Digital Ad Spend Climbed to $259 Billion in 2024,” writes Adweek’s Paul Hiebert. That’s up 15% from 2023 and is due in part to increased spending around the Summer Olympics and the presidential election.
• Another day, another new Substack. “Business Insider Founder Launches Regenerator, a New Publication,” on the popular newsletter platform, reports Adweek’s Mark Stenberg.
• For The Washington Post, Scott Nover writes, “Sarah Palin and the N.Y. Times are in court again. Here’s why.”
• And another, this one from The New York Times’ Katie Robertson: “Former Times Editor Testifies in Defamation Suit Filed by Sarah Palin.”

Have feedback or a tip? Email Poynter senior media writer Tom Jones at [email protected].

More resources for journalists

• Reimagine your reporting as a non-fiction book. Apply by April 25.
• Strengthen democracy through investigative journalism with our hands-on five-week seminar. Enroll now.
• Refine your immigration policy expertise with Poynter's Beat Academy. Enroll now.

Have feedback or a tip? Email Poynter senior media writer Tom Jones at [email protected].