Ar­ti­fi­cial In­tel­li­gence: Real Im­pli­ca­tions

Policy - - Before The Bell | From The Editor - BY DALE SMITH

While Mon­treal has emerged in the past five years as a global hub for ar­ti­fi­cial in­tel­li­gence (AI) re­search and tal­ent re­cruit­ment, Ot­tawa is where Cana­dian pol­icy in the rev­o­lu­tion­ary realm is be­ing gen­er­ated, and that makes it a hot topic for politi­cians, pub­lic ser­vants, jour­nal­ists and con­sul­tants in the cap­i­tal. In March of 2017, the Trudeau gov­ern­ment an­nounced $125 mil­lion in fund­ing for a Pan-Cana­dian Ar­ti­fi­cial In­tel­li­gence Strat­egy with a goal of mak­ing Canada a world leader in the field. And when the In­no­va­tion Su­per- clus­ters Ini­tia­tive un­veiled its five win­ning bids ear­lier this year, the SCALE AI su­per­clus­ter was among them. While AI — the repli­ca­tion of in­tel­li­gent out­come op­ti­miza­tion once the ex­clu­sive do­main of hu­mans now shared by ma­chines — has the po­ten­tial to of­fer undis­puted real-world ben­e­fits, there remain many pub­lic pol­icy, im­ple­men­ta­tion, and eth­i­cal ques­tions around the tech­nol­ogy. Be­fore the Bell hosts Catherine Clark and David Akin each hosted a panel of ex­perts and stake­hold­ers to dis­cuss those very ques­tions.

Dur­ing the Pulse seg­ment of the event, hosted by Akin, Chan­tal Bernier, coun­sel and head of Den­tons’ Cana­dian pri­vacy and cy­ber­se­cu­rity prac­tice, said that the cur­rent leg­is­la­tion may not be able to keep up with the con­sent im­pli­ca­tions for AI.

“The AI takes in data, for ex­am­ple my name, ad­dress, and pur­chase his­tory, and cre­ates a pro­file on me that I’ve never granted my con­sent for,” said Bernier. “The con­sent and trans­parency im­pli­ca­tions of ar­ti­fi­cial in­tel­li­gence would re­quire the mod­ern­iza­tion of the le­gal frame­work.”

Marc-Eti­enne Ouimette is di­rec­tor of pub­lic pol­icy and gov­ern­ment re­la­tions at El­e­ment AI, the Mon­treal com­pany co-founded by glob­ally rec­og­nized AI pi­o­neer Yoshua Ben­gio. Ouimette said that Canada has been ahead of the curve

when it comes to in­vest­ing in the de­vel­op­ment of AI.

“The story of the de­vel­op­ment of AI it­self is a mi­cro­cosm of this false di­vide be­tween fun­da­men­tal re­search and ap­plied re­search, and the need to fund re­search in the first place,” said Ouimette. “We wouldn’t have this AI break­through in Canada were it not for the fact that the gov­ern­ment in­vested over a twenty or thirty-year pe­riod into what led to the break­throughs in deep learn­ing and neu­ral net­works.”

Dan Duguay, prin­ci­pal at Tac­tix, said that there is a gap be­tween gov­ern­ment and in­dus­try based on the dif­fer­ence in level of un­der­stand­ing of where tech­nol­ogy is and any gov­ern­ment’s abil­ity to keep up with the head-spin­ning pace of in­no­va­tion.

“That’s a gap that’s dif­fi­cult to bridge, if in­dus­try and gov­ern­ment aren’t talk­ing the right way and un­der­stand­ing each other,” said Duguay. “The sec­ond prob­lem that AI is demon­strat­ing is the rate at which that tech­nol­ogy evolves and changes, and the rate at which gov­ern­ment stays on top of it. There’s an asyn­chro­nous na­ture to that which is even worse in AI.”

Duguay wor­ried that the gap may be­come un­man­age­able with­out prin­ci­ples-based leg­is­la­tion.

Dur­ing the Pol­icy seg­ment hosted by Clark, Sigfried Usal, manag­ing di­rec­tor of cortAIx at the Thales’ Cen­tre of Re­search and Tech­nol­ogy in Ar­ti­fi­cial In­tel­li­gence ex­per­tise in Mon­treal, said that the greater con­nec­tiv­ity of sys­tems is pro­duc­ing a lot more data than it used to.

“You have to deal with that mas­sive data

The prob­lem that AI is demon­strat­ing is the rate at which that tech­nol­ogy evolves and changes, and the rate at which gov­ern­ment stays on top of it. There’s an asyn­chro­nous na­ture to that which is even worse in AI.” — Dan Duguay Prin­ci­pal at Tac­tix

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.