Five years ago, the conversations around AI were decidedly dark. It wasn’t hard to see how tech like facial and pattern recognition could, say, aid an authoritarian regime. Or how flawed data could skew machine learning outcomes. But from the prophesied ashes of humanity bloomed a movement to encourage all people dealing in AI, from data scientists to biz dev execs, to focus on ethics and helping each other. “It’s an initiative aimed at establishing a more balanced dialogue around AI,” says Amir Banifatemi, co-founder of the AI for Good Global summit. “In particular about how humans and machines can collaborate to build a better future.”
Banifatemi is also GM and Chief Innovation and Growth officer for the XPRIZE Foundation, a non-profit on a mission to accelerate the development of new tech that benefits humanity—through big-money competitions. Because nothing turbocharges innovation like a top-quality contest. In fact, some of our greatest advancements came about because someone put up serious cash. Canned food? Thank Napoleon Bonaparte, who offered thousands of francs to better feed his sprawling army. The first solo trans-Atlantic flight? A $25K prize launched the aviation industry, attracting millions in investment, jumpstarting technological breakthroughs, and capturing the world’s imagination. Now, thanks to a similarly audacious prize, artificial intelligence has burst through Uncanny Valley tropes—and worse—to become an indispensable friend to humanity. “Amir and this prize really kickstarted the whole AI for Good movement,” says Neama Dadkhahnikoo, XPRIZE’s Director of AI and Data Operations.
The IBM Watson AI XPRIZE launched in 2016, offering $5 million to any team who could revolutionize AI’s place in society, proving its use as an essential force for good. Create AI technology that has a huge and measurable impact on the world at large, and you win.
Over 150 teams answered the call, developing products in a wide range of applications, from recycling to public health, education to medicine. Then at the end of 2019, an independent panel of experts including academics, NGO professionals, and venture capitalists with deep AI expertise, narrowed the competition from 34 qualified teams to 10 semi-finalists—with a little machine help, of course.
“We ran a concept called a Monte Carlo simulation,” Dadkhahnikoo says. “Which is a fancy way of saying we ran a bunch of simulations to see how the biases of the judges would affect the rankings.” A judge with a medical degree, for instance, might lean toward health applications. Each judge would assess a handful of teams, then machines crunched their scores to help XPRIZE develop a method for finding a true, ideally unbiased ranking.
The judges assessed four main criteria: achieved technical impact, evidenced real-world impact, scalability, and ethics. Note the words achieved and evidenced. The XPRIZE Foundation does not deal in hypotheticals; they demand results.
To that end, IBM Watson was the perfect partner. The computing juggernaut’s suite of AI tools would help competitors more quickly realize and scale their products, sharing processing power and reams of data necessary to animate certain visions. The XPRIZE Foundation brought expertise in contest design, having launched nearly $300 million in prizes over the last 25 years.
Each of the 10 semifinalists presented their solutions to the judges and a live audience at TED HQ in New York City just before the pandemic, and three teams moved on to the final round. Those finalists are tackling the challenges of mental healthcare, human trafficking, and malaria, respectively, and are vying for the winning prize of $3 million, with $1 million going to second place, and $500K to third.