The IBM Watson AI XPrize was announced earlier this summer (H/T TechCrunch). Teams have until December 1st to register (early bird deadline is October 15) for the four-year competition, which concludes at the April 2020 TED conference. The goal is a bit more open-ended compared to other XPrizes, in part to accommodate potential changes in AI technologies and capabilities. This explains the ‘wild card’ portion of the competition, where companies who did not submit at the beginning of the competition have the chance to develop proposals for consideration in later years.
Teams must develop a four-year plan for applying AI to a grand challenge, including milestones, testing processes, and overall solution. The plans must have some description of the anticipated AI technologies involved as well as how humans would work with these technologies in addressing the grand challenge. Each year, a team will submit a report and testing results. Should they wish to be considered for that year’s Milestone Awards (as defined in their plan), the team will need to apply for a spot at the annual IBM World of Watson. Of the (up to) 10 teams accepted, two will be recognized with Milestone Awards, and some teams will be eliminated from further consideration.
In the third year of the competition, the field will be reduced to three teams, and these teams will give a TED Talk at TED 2020 prior to the final prizes being awarded. Over the four years of competition, a total of $5 million will be awarded. Most of that will go to the three finalists ($3 million for first, $1 million for second and $500,000 for third) with the remainder distributed for Milestone prizes.
Again, teams have until December 1 to register, with competition ramping up early next year. The involvement of IBM Watson makes sense. Watson is the computer best known for competing on Jeopardy! a few years ago, but it has made strides in artificial intelligence since then, with IBM working to provide services for those seeking to process large amounts of unstructured data in ways that more closely resemble human thinking. While I don’t see any indication that IBM is looking to integrate Watson into the competition like Intel has incorporated its technology into some of its competitions, I wouldn’t rule it out.