Celebrating the tenth-anniversary winners of the AI4Good Lab

What happens when you give young people the skills and training to create AI tools for social good?

You get a world where food banks don’t run out of food, period products are always within reach, dangerous falls are prevented and online manipulation is exposed.

These were the projects that won this year’s AI4Good Lab Accelerator Awards. 

A decade of purpose-driven innovation

The AI4Good Lab is a seven-week program designed to empower women and gender-diverse individuals with skills and mentorship to build their own machine learning tools. Co-founded by Doina Precup, Canada CIFAR AI Chair at Mila and associate professor at McGill, and Angelique Mannella, their goal was to address gender inequality in the AI industry. 

Today, that mission remains more important than ever. Statistics show that in Canada, women only make up about 22 per cent of tech workers and on average earn $20,000 less than men. In addition, the 2026 Stanford AI Report put female representation in AI in Canada at 30 per cent. 

“Offering an inclusive learning environment that is also free of cost and provides a stipend for participation is really necessary to open doors,” explained the Lab’s Program Manager, Jennifer Addison.

Operating across the country, the program is led by Mila in partnership with CIFAR, the Vector Institute and Amii, welcoming trainees in Montreal, Toronto, Edmonton and virtually across Canada.

Meet the winners

In the Virtual stream, Arzow Maksum and her team built Perspect, a free Chrome extension that automatically scans text on a user’s screen to identify manipulative techniques. 

Powered by natural language processing and machine learning models, the tool detects and explains hidden persuasion on news sites and social media platforms like Facebook and Instagram.

While Maksum is a software engineering student, she explained that the diversity of her cohort was partly what made each project unique.

“You don’t need to be the most technical person to enter the Lab. There were so many people from diverse backgrounds — law, chemistry, graphic design. The Lab is well-equipped for technical students, but they’re also very cognizant of those coming from non-technical backgrounds.”

In Toronto, one team leveraged their diverse skill set to develop the award-winning project, OKAI Period. 

As one in four menstruating people in Canada face period poverty, they built a spatial map of free period products across the city. Using smart mapping algorithms and public data, the platform highlights underserved neighbourhoods and predicts where shelter and community needs will peak next, ensuring non-profits can put resources directly in reach.

“It was the best seven weeks of my life,” said trainee Mahnoor Butt. “We formed this sisterhood, this community, where we were all there to help each other.”

Sophia Ramotar, part of the Montreal cohort, had a similar experience working with her team on their project, ARIA.

They built a mobile app to address freezing of gait in Parkinson’s patients, which frequently leads to dangerous falls and injuries. Their solution builds on Rhythmic Auditory Stimulation (RAS), a clinically proven technique that uses audio cues to improve stride length and cadence. Unlike traditional RAS, ARIA predicts freezing episodes seconds ahead, automatically delivering adaptive beats without clinical supervision.

Before joining AI4Good Lab, Sophia struggled with self-doubt. 

“The job market is a little bit daunting for a lot of people in computer science right now…It’s very competitive. But being in a larger group of women who are working together was a really uplifting environment,” she shared. 

Replacing pressure with collaboration was a deliberate choice for the lab.

“We try hard to dissipate competitive energy,” Addison explained. “They’re all extremely smart and come from a university environment where they’re graded on everything. Asking them to unlearn that mindset can be hard, but it’s essential.”

This intention was felt by the Edmonton team behind FEEDS, who were in disbelief when they won. 

“It was definitely a surreal moment,” they recalled. “We learned so many lessons, made friends and had so much fun along the way. That’s what we’re forever grateful for, not just the fact that we actually won.”

FEEDS was created to support food banks on a provincial and regional level. With visits nearly doubling since 2019 and grocery prices outstripping inflation, food banks struggle with unpredictable demand. 

To help staff plan ahead, the team combined regression and time-series forecasting techniques, drawing on indicators and historical records from regional food banks to project supply and visitor trends. This allows food banks to optimize inventory planning well before shelves go bare.

Looking ahead

These four projects are just a few examples of the important work the AI4Good Lab has produced since its creation. Looking ahead, the Lab’s Curriculum Manager, Yosra Kazemi, is confident the program will continue adapting to shape the future of the field.

“Over the past decade, the Lab has evolved as AI has evolved. But there is still so much opportunity ahead to keep growing and help shape a generation of people who are not only able to build AI, but are thoughtful about why and for whom they are building it.”