Research by Topics

Can AI be better trained? New research suggests yes

May, 2023 | Article

A new study reveals that training AI systems with normative labels rather than factual ones significantly improves their accuracy in predicting human normative judgments.

Mitigating risks & maximizing benefits: How AI can be regulated in the Canadian financial sector

Apr, 2023 | Article

As AI transforms the financial sector, Canadian regulators are faced with the challenge of balancing innovation with consumer protection. The right regulations might already exist.

Why the proposed AI development pause may do more harm than good

Apr, 2023 | Article

Six months probably isn't enough time to create meaningful guardrails for the technology, and a ban on development may even hinder our ability to solve some of the deeper issues with AI.

'Don't fire the humans.' Why people are still needed through AI's boom

Mar, 2023 | Article

As artificial intelligence gains ground, guardrails are still needed, new research finds. Pairing AI with human intuition will create better outcomes for businesses.

5 myths about Gender Analytics

Feb, 2023 | Article

It's tempting to think that an inclusive approach to analytics is just an add-on to a day-to-day job, or worse, isn't profitable. But those are all just myths, say GATE's Sarah Kaplan and Lechin Lu.

AI could fundamentally change the way systems are built...if we let it

Nov, 2022 | Book

Everything is built on systems of constraint, and until we revisit the rules that glue these systems together, AI will only ever add incremental value.

Management analytics explained

Oct, 2022 | Article

What exactly is management analytics, and why should businesses buy in? Professor Opher Baron shares his thoughts.

Can AI address racial disparity in pricing?

Feb, 2022 | Article

Race-blind algorithms helped decrease the revenue gap between Black and white Airbnb hosts by more than 70 per cent. But recent research from professor Nitin Mehta finds AI isn't a panacea to addressing these inherent biases.