AI and Robotics at TRU: Institutional Augmentation – Part 3

“The basis for fear can actually be overcome by understanding,” TRU president Airini said. “It is a co-pilot that will help us in our teaching, in our learning and in our research in remarkable ways that will help us to move from the worries around artificial intelligence to instead see this as augmented intelligence. The human agency here is vitally important as we move forward.”

TRU brass and faculty spoke to CFJC on the subjects of ethics and vision for Episode 3 of AI and Robotics at TRU: Institutional Augmentation.

“Now, it’s not just limited to engineers who can code, who can leverage AI. All of us can,” said Li, noting 2023 is when Chat GPT entered public domain. “When technology changes that fast, it is the responsibility of institutions like a university to educate the community on how you use certain things responsibly and how do you define responsible.”

Quan Nguyen, a TRU assistant professor in the computing science department, is a proponent of AI but sees danger in overreliance.

He spoke to CFJC in the TRU Collaboratory, in which there are 10 super computers and a sign that reads, ‘Learn more about generative AI.’

“Part of learning is productive struggle,” Nguyen said. “But, if there is no struggle, if everything just comes so easy, instant gratification, then there is no learning and we’re becoming more dependent on technologies.

“In a typical homework assignment, students could copy and paste the questions into Chat GPT and get the solution key. By doing that, they short circuit the thinking process.”

Nguyen, along with student research assistants, developed an AI assistant, TRU Think, designed to augment human learning through Socratic methods.

“The system will take the questions and break it down into smaller steps and ask the students to reflect on what they think they should do next based on the hints,” Nguyen said.

Zeinab Teimoori, an assistant teaching professor in the engineering department, cautions her students on the perils of irresponsible AI use.

“Don’t rely 100 per cent on AI answers,” she said. “After you get the result, it needs validation. If you do not validate, it means that you are relying 100 per cent on one mind that was generated by humans. That mind could be biased.”

Added Li: “It’s very easy to fall in the trap of just believing it. AI is known to hallucinate a lot and make things up because AI in the background is just a bunch of statistics with some magic of randomization. It could give us misinformation which allows us to make bad or wrong decisions. When we think about responsibility and ethics, a lot of it is based on transparency – how we create the right system so we keep a human in the loop.”

Teimoori conducts workshops on campus to introduce AI concepts, including for Grade 7 students.

In an advanced software lab, she shows them how to create customized generative stories using source code, emojis and keywords.

“When we were hearing those stories, we realized how creative these young minds could be, some amazing results from their own AI they generated within one hour,” Teimoori said.

“We need to be very, very careful what we are feeding to these young minds and probably we can train them to save the world in the future and not let AI take control of the earth. “

Teimoori smiled while making the comment on AI world dominance, but the existential threat of AI developing its own AI which seeks to destroy humanity – a subject for science fiction flicks – is fodder for ethics questions.

“The large quadruped I was working with in Italy, if it were to kick you, its foot would go right through your leg,” said TRU robotics expert Geoff Fink, an associate professor in the engineering department. “The bones would be nothing to it.”

Fink is not sold on robots taking over the world any time soon.

“Because it is exponential growth in intelligence, that is where people get a little bit worried,” he said. “Yeah, we can control it now, but what happens if it’s a million times smarter? I’m still not too worried. If all of the robots in the world were to unite against us right now, it’s nothing compared to any kind of military force.”

AI taking human jobs is no science fiction. Fink said it takes jobs in some industries and creates positions in others.

“I’m still fairly certain the net number of jobs stays about the same,” he said. “I’d like to stress the majority of these robots we like to put in situations that are dangerous for people.”

Fink sees opportunities in areas such as accessibility.

“If you have somebody in a wheelchair, for example, they wouldn’t be able to plant a tree on a mountain, but they can control a drone that is planting trees on mountains,” Fink said. “It opens up more people in the labour market to help in some of these causes.”

Teimoori said her job is an example of one that appears to be safe.

“It would be a little difficult to just rely 100 per cent on a robot to come to a classroom and teach students because it still needs interaction,” she said. “Humans have emotions. Robots don’t. Also, critical thinking and being creative, these are skills humans have and robots don’t.”

Added Airini: “It has no emotions. It is actually a prediction machine for us based on the constraints we put in. We say, ‘Of all the pieces of information you could draw on across the world, have a look at these ones and produce this kind of report, with bullet points or in poetic form or in Italian.’”

And, what if robots are eventually equipped with emotions?

“I hope I’m not in that era,” Teimoori said with a laugh.

In his course Internet of Things, TRU assistant teaching professor Anthony Aighobahi helps students apply AI to objects.

“Security is a big factor because of the cyber threats we face these days,” he said. “Many of these systems, if security is not considered, it becomes a tool for attackers.”

AI is part of the business plan at TRU.

“How do we create the right type of partnerships across Canada, so that we’re able to put all the innovators who are pushing the envelope in their space together?” Li said. “Our firm belief is collaboration is what’s really going to beef up Canada’s capabilities across AI and beyond. There is a lot of conversation around Canada’s AI sovereignty, as well.”

One of those partnerships is under construction, with BELL AI Fabric network infrastructure coming soon to TRU campus.

Data centres are proving harmful to the environment, but Airini said TRU is committed to sustainability, noting its low carbon district energy system will repurpose energy from the Bell AI Fabric infrastructure.

“Later this year, when that goes live and when we have the AI infrastructure here, we will be carbon zero,” Airini said.

Added Li: “We’re thinking through how we can create these new-age data centres that are significantly more sustainable than the ones that are in the U.S. that the large cloud players had previously.”

Airini said TRU is aiming to be one of Canada’s leading universities in innovation.

“The ability to create something new, to patent it and to monetize it is important and we can see that possibility in the AI space,” she said.

Li said she is working on a pan-institutional program – TRU Horaizon – that will promote equity, strengthen the community and leverage principled AI in use-inspired research.

“Our ambition is to bring more folks from the community in once we have the right framework set up,” Li said.

“One of our pillars is game changers, research clusters that are really pushing the envelope on creating first-of-a-kind products or research in areas that are completely built on top of AI.”

AI is the most disruptive and fascinating technology in human history, according to Nguyen.

“There is going to be more capability for it to interact with the physical world, for it to understand you’re sitting next to me, to understand you’re looking at me, for it to understand my facial expression,” he said.

“Where is the limit and what can we do in the future? We really don’t know. What we can do is stay on top of development and think strategically how to use for our own benefit and in a way that aligns with our values.”

Next week, in the fourth-and-final instalment of the series, health experts at TRU will demonstrate AI tools in nursing, stroke monitoring and cancer classification.

“Our world is one in which AI is rapidly becoming as present as air and it matters that this university is at the leading edge for this work because we see it as important in how we’re able to serve our region and the province,” Airini said.

by Marty Hastings

AI and Robotics at TRU: Institutional Augmentation – Part 3

Our Community