Dal starts new cross-faculty Atlantic Canada AI Institute

Dalhousie University has started the Atlantic Canada AI Institute to be the region’s hub for AI development with a focus on responsible use and cross-discipline research.

The institute, called A2I2, launched in August and currently has around 50 faculty and over 100 students involved. 

The institute draws from fields of computer science, engineering, health, medicine, social science, law, management, science and economics, says Frank Rudzicz, the institute’s director. All of the institute’s “core members” listed on its website are in the faculty of computer science.

“We want to see [AI] develop locally and have people who live and work in research here be part of the conversation,” says Rudzicz. “As opposed to just using tools that are developed elsewhere.”

Currently, the institute’s members are using AI to analyze brain data, build a risk dashboard for post-cardiac surgery patients and monitor livestock and agricultural equipment. They are also researching how AI and music can “coexist peacefully,” among other projects. 

“Everyone has their own perspectives, so we wanted to expand who is at the table,” says Rudzicz on the cross-discipline research focus. “[The models] are way more powerful than they’ve ever been; they’re only going to get more powerful, and we need to be really mindful about how we deal with that.”

“One thread we’re trying to weave through the whole institute is the idea of responsible AI,” says Rudzicz. He says much of the institute’s current work focuses on AI safety, including how to safely manage AI-training datasets while protecting personal information.

Rudzicz says the institute wants to start working on more top-down, large-scale projects once they formalize their internal governance structure later this fall. Other future goals involve more collaborations with local industry, holding events, community engagement and working with the government on potential AI legislation.

Iaian Archibald is the executive director of DeepSense, a Dal-based organization created by the faculty of computer science to connect students with the industry and develop AI use in Atlantic Canada.

Archibald says that while Canada’s AI industry is surging, Atlantic Canada is “very much behind.”

He says an AI institute like Dal’s is necessary to promote long-term interdisciplinary work now that AI has spread across society, as well as to capitalize on industry interest in Nova Scotia and the region.

“We’ve got incredible under-leveraged talent [in Nova Scotia]. We’re graduating some really, really high-quality individuals and a lot of them want to stay here,” says Archibald.

“We need to do a better job connecting this talent with industry, and there’s a huge opportunity there for Dalhousie and the AI industry,” he says. “Dalhousie is at the centre of all this.”

The new institute developed out of Dal’s previous AI research as the successor to Dal’s Institute for Big Data Analytics, which was founded in 2013. That institute was reworked into its new form after the senate approved the change this year. Rudzicz says the change was spurred by the explosive development of AI in the past decade, which required an institute with a broader scope and scale.

Along with cross-faculty collaboration, Rudzicz says another goal of the institute is to include many student perspectives, including anti-AI and AI-skeptical ones.

“The main thing is not to discount the fears … myself and a lot of people who’ve been working [in AI] for a long time also share the exact same fears. It’s not an us-versus-them scenario,” he says. “What we envisioned was more R2-D2 and less Terminator.”