How IBM’s Project Debater Could Fix Facebook

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Much of what we currently get on long-form social media is abhorrent. (Twitter is short-form social media, and it has its own unique issues.) It is either questionable news, arguments started and maintained by trolls, or updates about subjects that most of us couldn’t care less about. I believe this is likely the real reason why Facebook growth has stalled.

But what if you could make social media relevant or even interesting?

Ever since I was briefed on IBM’s Project Debater, I’ve been intrigued with the concept of an AI-driven, crowd-sourced debate engine. Applied to a social network like Facebook, this technology could result in some well founded and far more interesting discussions. More importantly, it could result in far less embarrassment and a higher probability that citizens would become better informed and more engaged in the election process both locally and nationally.

Project Debater

Project Debater is a follow-on project to the Watson Jeopardy challenge. It is a separate development effort leveraging Watson, somewhat like the Jeopardy project initially was. Debater is very different than anything currently in market. The project uses crowd-sourced arguments on a selected topic, and then the participants vote on the best argument consistent with their respective positions. So let’s say the argument was on the topic of the US space program and whether it should be largely privately funded or, as it once was, mostly funded by the government.

People would pick a side, submit arguments, and then vote on which argument was strongest. Then the two strongest sets of arguments would be pitted against each other, and a winner would be determined. The underlying AI would learn the process, learn how to select the best argument, and assist with fact-checking to assure the arguments were founded in fact.

This not only forces rigor into the discussion but helps assure that people are arguing with real facts and not something that they made up to win the argument.

Forums and Facebook

The problem with long-form social media like Facebook and forums (and I’m a moderator myself) is that too much of what is there is trivial and too many discussions degrade into name-calling. The posts are either uninteresting or aggravating, with no good middle ground. But what if you auto-curated the content with something like Project Debater and scored the responses in real-time based on information accuracy, clarity of argument, and how compelling the argument was? And then rolled that up into a score.

With that technology in place, people posting would take time finding accurate links (assuring information accuracy), learning to be brief and to the point (clarity of argument), and focusing on arguments that changed minds (how compelling the argument is).

You would get discussions that are increasingly more interesting, and you would be more likely to learn something from the process — other than new words you can’t use in polite company. And over time, people would learn a set of skills that would allow them to win arguments with real facts and preparation, not just voice volume and stubbornness.

You could even rank those that were participating regularly and give titles like “Master Debater” to those that consistently demonstrated high quality and win the debates. And for some career paths, such as attorney, contract negations, and even parent, the improvement in argument skills could have a very positive impact on the quality of the job and the upward mobility of the employee.

We live in a hostile world, but much of the hostility, I believe, is the result of people being tricked into taking positions that are both not based on fact and are against their own best interests. Social media has been a huge liability in this process, but I don’t think it needs to be. If we can raise the quality of discourse, then social media, and both internal and external forums, could be used to increase discussion skills so that disagreements could focus on the merits of each argument and less on the emotions surrounding them.

This is all very consistent with IBM’s strategy to develop AI capability that will enhance humans. By flipping discussions to facts and away from name-calling and poorly founded emotional arguments, they help move the ball forward in terms of making work less hostile, and positive progress more assured. Of course, in the hands of a politician or attorney, the IBM Project Debater tool could become that killer app that assures a win, but until then, it could still make us all better at making compelling arguments.

Now if our kids get hold of a tool like this, we are pretty much done for.

Photo courtesy of Shutterstock.