But the point was never that the moderators alone must wield review power. That power could be shared, given the stakes. There may be a way. For that, ABC News must be willing to do the following:

Principle 1: Collaborate. Create a crowdsourced partnership for reviewing claims live.

Principle 2: Design in some inconvenience and friction. That means having a fact-checking round for every topic, at the end of the next topic. It is a small cost to favor democracy.

Say there are four topics: Economy, Abortion, Immigration, and Foreign Policy. The same rules that CNN used for the first debate can be adapted to insert one deferred fact-checking round per topic.

First, well ahead of the debate, bring together at least 10-20 external fact-checking teams in a loose collaboration with the show’s producers. These are teams that are already going to be doing fact-checking for their own outlets and their social media audiences. Instantiate a short-term targeted partnership only during the presidential debate and disband after that. The Duke Reporters Lab documents nearly 80 fact-checking units in the U.S., both inside and outside of news entities. The U.S. is a major base for fact-checking and a leader in capacity building and training for others. There is no shortage of talent.

There are plenty of digital workflows and tools that let teams collaborate securely in real time in the public interest on a claim-by-claim basis. Invited partners would need to have teams available to simultaneously spot and review claims as they are made live, share notes quickly, and submit initial findings within 5-15 minutes, while the debate is going on. A diverse group of partners will be able to pick and analyze a diverse set of claims they already have knowledge about. Some might pick gender and abortion. Others, economy and jobs. Still others, Ukraine and Russia.

The partnership’s role is to submit a list of dubious or clearly false claims to the producer’s team with accompanying context, as the debate goes on. Each submitted status for a claim made by either candidate could have the following basic data:

  1. Is this a new claim altogether or one that the candidate has already made?
  2. If not new, has this claim already been debunked, if so by whom, when, and with what counter-evidence?
  3. If totally new, what can be said about any trustworthy public evidence supporting or debunking it?
  4. Has any news organization already asked the candidate or their campaign to submit their evidence for this claim and did they get a reply?

As each segment progresses, the producer’s end will see a flow of claims and analyses coming in from the partners. They looks at every dubious claim and elevate the most questionable, or those with the least evidence, or those already debunked, or the already known lies, to the top of the list. The more categorical the finding by two or more fact-checkers, the more likely it goes to the top of the list, compared to other claims.

It then becomes the producer’s prerogative to let the top claims on this list flow to the moderators in a predetermined format co-designed with the moderators. Each item on this list will have the false or incorrect claim or lie, the topic (or debate segment), which candidate said it, and a summary of what’s been found out by partners. In addition, the producer’s team could add a question to pose to the candidate. This becomes the setup for the moderators’ fact-checking or challenge round.

Imagine the debate completed segment one, Economy, and moved to segment two, Abortion. By the time the debate is midway in the Abortion segment, the debate’s producers ought to have started seeing a list of dubious claims (some will be lies) with supporting context from the partners. When the Abortion segment finishes, the moderators announce the fact-checking round on Economy. Likewise when the Immigration segment completes, they run the fact-checking around on Abortion.

In each fact-checking round, one false claim at a time is offered with evidence and a question to the candidate who made it. After the candidate has responded, the competing candidate gets a shot. The candidates don’t have to agree with the moderators, but the public will hear what the moderators are putting on the record. And, there needs to be proportionality. If one candidate made 10 false claims or lies and the other made two, the split of claims for that segment needs to reflect that. The moderators must state those numbers openly.

Take the four segments. Allocate 12 minutes per topic and 8 minutes of fact checking. That is 80 minutes for four topics, leaving 10 minutes for moderators and breaks. When the fourth topic ends, you’ll have two fact-checking segments back to back. There is a way to squeeze in closing statements too. Undoubtedly, there will be some friction between moderators and candidates. But as long the claims are litigated around evidence rather than just aired, it will be worth it.

At each fact-checking segment, the moderators could mention that this is part of a collaboration. The value of partnered journalism will be evident to millions of people.

The point is this. Even if a collaboration can produce a few lies or false claims per segment and the moderators can use that to hold candidates accountable, it will change the tone and content of the debate. The candidate who spouts more lies ought to get called out more.

Completely separate from fact-checking rounds in the main TV debate, ABC News could go further and offer a second screen on the web, curating the vetting of claims done by the partners as a stream. This will help a chunk of viewers watch both the debate and the fact-checking stream from the same brand, at the same time.

TV networks can and must do something better than placing raw claims on the air and expect the presidential candidates to fact-check their opponents during their rebuttals.

Thanks to Sara Catania, Chief Program and Operating Officer, Solutions Journalism Network for reviewing this idea and offering a suggestion for a second fact-checking screen.