Fake news, prepare to be exposed.
And if you’d like to put the day of reckoning in a calendar, Sept. 10 would be the relevant date.
That’s the day the winner of the $1-million Leaders Prize will be announced, and the day that the scourge of fake news will officially have a new, AI-driven nemesis.
Communitech, in partnership with the Schulich Foundation and the Leaders Fund, along with the University of Waterloo as academic partner, launched the Leaders Prize in November of 2018, a first-of-its kind, all-Canadian competition designed to encourage teams with expertise in artificial intelligence to solve a global problem of significance. The specific problem – fake news – was announced the following June at True North 2019, Communitech’s signature tech-for-good event.
The $1-million Leaders Prize was to be awarded at this year’s True North Festival, which was slated to unfold June 1-7. But the award was put on hold along with the True North Festival itself, due to the COVID-19 pandemic.
Ten teams were selected last January from 150 applicants, and one has since dropped out. On July 17, the nine remaining teams will submit their Phase 2 solutions. The top prize in September will be awarded to the team with the “most effective and efficient AI-based fact-checking solution that exceeds 75 per cent of the score achieved by a human baseline.”
“The top nine teams have a huge opportunity to solve a major worldwide problem with broad applicability and to inspire the next generation of Canadians to pursue a career in technology," said David Stein, co-founder and Managing Partner, Leaders Fund. “We are excited to see their solutions, and believe they will effectively combat fake news, helping to restore trust, strengthen democracy and reduce the fraud affecting the health of citizens worldwide.”
The emergence of COVID-19 has had a significant impact on the competition beyond delaying the awarding of the prize. Fake news has proliferated during the pandemic.
“What COVID has changed is my perspective on how difficult the problem of fake news is,” said Toronto’s David Brait, of the team dbrait2000. “The daily deluge of COVID information has evolved so quickly and been so tough to parse for even the most qualified people. It has created such a difficult environment for platforms, consumers and experts to discern what is real and what isn’t.
“Such a rapidly evolving situation with such high stakes has shown why we need to get a better handle on fake news.”
Brait’s observation was echoed by Toronto’s Nima Shahbazi, of the team Nima-Fit.
“The pandemic has had a big impact on my modelling,” said Shahbazi. “Most importantly, the date and time that the claim is made needs to be taken into account. With other subject matter, ‘true’ or ‘false’ classifications are pretty static, but the truth about COVID-19 changes every day. It’s essential that the model checks trusted sources like the WHO and CDC constantly and takes the sequence and timeline of the updates into account.”
The $1-million prize will be awarded during an online event. Other teams involved in the competition include: Jaymody, Supersonic, NL Philosophers, Traveling Wilburys, ImAnLearn, Ava, and ConvolutedLP.