How AI Could Make Political Debates Smarter

Smart candidates is a good place to start, but AI could provide interesting fact checking and audience reaction in real time. JL

Erica Schroeder reports in Ad Week:

Humans need help, not replacement. It isn’t a moral question about firing humans in favor of machines. It’s an issue about information and insight, and who’s better at it. When computers replace people, they often fail at simple tasks that humans take for granted: pattern recognition, strategy, critical thinking, empathy, creativity and so on. AI systems can manage to see patterns in vast data sets. Fact-checking could become even instantaneous. Moderators could rely on data dashboards for their role. An AI-based dashboard could provide instant audience feedback.

The last presidential debate was a happier outlier in this sad election season.

Kristen Welker’s valiant work keeping the proceedings on track is appropriately being credited with redeeming debates as an opportunity to see candidates joust over substantial issues. Sadly, she spent energy and brainpower on keeping order, not delving deeper into substantive, relevant topics that citizens want to hear from the candidates on.

Even with the conclusion that the proceedings improved, dispassionate, unemotional machines can do better than the poor human journalists who suffer and soldier through today’s political debates.

Let’s look at the grim reality of rest of we saw the last time the candidates went head-to-head. The first presidential debate frustrated Americans for incivility, interruptions and flippancy. The dialogue was shallow, chaotic and mostly irrelevant to the issues voters care about. No matter our politics, we all agreed it was unbecoming behavior in the process to determine the highest office in the land.

That debate underscored what many Americans dislike about election debates: They’re boring, they lack substance, the candidates evade questions or use rhetorical tricks to deflect criticism, change the subject or shift responsibility when it suits their needs—and when candidates do speak clearly, most voters feel as if their concerns are ignored or glossed over.

The inevitable call for AI

Humans need help, not replacement.

It isn’t a moral question about firing humans in favor of machines. It’s an issue about information and insight, and who’s better at it. When computers replace people, they often fail at simple tasks that humans take for granted: pattern recognition, strategy, critical thinking, empathy, creativity and so on. However, AI systems can manage to see patterns in vast data sets, something we struggle with.

Don’t put robots in charge of debates, but use the concept of AI to help communicate with insight. How could AI improve the jobs of Kristen Welker, Chris Wallace, Susan Page and election debate moderators in the future? How could it have helped Wallace and Page turn in a performance more like Welker’s?

AI could empower journalists in the same way it helps doctors, bankers, investigators, retailers, scientists and others by granting them enormous powers to analyze and understand the data used in their daily work. It could help them use not only the information they access directly even better but also leverage enormous real-time and historical data sets to find new patterns or anomalies.

Artificial systems could do the same job without a shred of moderator bias that audiences see (accurately or not) in the debates. For this last 2020 debate, the debate commission had already adopted the ability to mute the candidates. Welker added flexibility to enforce or loosen the leash where appropriate.

AI could go much farther than that, and positively enhance the experience for the audience and the moderators. Fact-checking could become even instantaneous. Moderators could rely on data dashboards for their role. An AI-based dashboard could provide instant audience feedback.

The AI could help keep the debate producers and the moderators be more honest and objective, too. Follow-up questions could better represent real voter interests. AI could analyze voter social media comments or pundit insights, or the candidates’ public remarks (or lack thereof).

Welker did well to return the debate to more of what we expect. But moderators need to be taken out of their comfort zone, too.

Critics will say that AI creates new unforeseen problems. But research also shows that AI doesn’t need to harm processes where it is applied. Instead, it can amplify the rate and degree of positive innovation in a system. Again, it’s about providing debate producers and moderators with an additional resource, which they can exercise at their disposal.

The takeaway is clear: the debates—and our politics—need help. If we want to preserve our democracy, we should get on with it.