No Debate: Watson is revolutionizing the 2020 US Open

By | 2 minute read | August 28, 2020

At my family dinner table growing up, debate was the main course. Perhaps debate is too kind a word. My father, brother and I would argue. Passionately. Emotionally. And on rare occasions, logically. About anything, really. Politics, religion, current events. Nothing was off the table.

But the thing we argued about the most was sports. Dave Winfield or Dave Kingman? Joe Namath or Johnny Unitas? And for a long time I thought our little dinner tradition was unique to the Syken family. Of course, I later learned that all sports fans engage in similar debate, at dinner tables and in dorm rooms, at bus stops and in bars, on sidewalks and in stadiums, all around the world.

And this is the world in which we are debuting Open Questions with Watson Discovery, an AI-powered engine that is curating tennis debates for the 2020 US Open. Throughout the tournament, on USOpen.org/openquestions, fans can contribute their opinions to a series of tennis debate topics, from “Is Roger Federer the best men’s tennis player ever?” to “Was Chris Evert vs. Martina Navratilova the greatest tennis rivalry of all time?”

But this is no dinner table conversation. Watson has used powerful natural language processing to analyze 14 million data sources around each topic – articles, blogs, transcripts, and more. It identified both pro and con arguments for each debate, and even assessed the quality of those arguments. When fans land on the site, they will find a summary of the debate, grounded in facts. But they will also be able to influence Watson’s analysis by adding their own voice to the debate.

Why use a powerful business AI to facilitate tennis debates? Two reasons. First, the USTA was looking to foster conversation and community for tennis fans in a year when they could not attend the US Open in person. Along with Match Insights with Watson – which uses many of the same capabilities to create AI-driven cheat sheets for each match – we think Open Questions will drive new levels of fan engagement. (By the way, both features were conceived, developed and delivered in just 12 weeks.)

Second, we are at another inflection point in IBM Research’s development of AI, and we are eager to share it with the world. More than a decade ago, I was part of the team that put Watson on the game show Jeopardy!, which marked the end of the AI Winter and a milestone achievement in natural language processing. Together, Open Questions and Match Insights represent another step-change in the evolution of AI, because they are not only able to understand natural language, but they are also able to generate it, creating cogent narratives from both structured and unstructured data.

We had another tradition growing up in my house. Every year, we would measure our height with a notch on a doorframe, the date and age written in pencil beside it. But some years, in our early teens, we grew so fast we had to measure our progress every month. I believe Watson is experiencing a similar growth spurt this year, at a time when we need it the most. From the real-time insights Watson Discovery delivered at the GRAMMYs back in January, to the life-saving information Watson Assistant has been delivering to countless citizens throughout the COVID-19 pandemic, AI is revolutionizing the way we entertain, engage, and educate.

Not even my family can debate that.