We all have our opinions on AI; it’s time to ask the experts
Yasmine Elsayed (she/her) // Contributor
Rachel Lu (she/her) // Crew Illustrator
Ren Zhang (they/them) // Illustrator
AI (artificial intelligence) is advancing at a pace that challenges many things to do with how people learn, work and even interact with one another. The incorporation of AI in our daily lives might appear to have happened overnight; however, AI has been in the works for years. The topic of AI has been brought up countless times, especially in education. Institutions often struggle to keep up with rapidly changing technology, and the divide between what students are trying to learn versus what the educators are trying to teach can be taxing. Many educators introduce it as a helpful tool, while others refuse its use in their classrooms, but few educators take the direct approach to closing the gap like Jason Madar.
As a former software developer at IBM and a CTO for many tech startups—now a computer science professor at Capilano University and a program coordinator at Langara College—Madar brings his field experience into the classroom, where he focuses on preparing students for a digital landscape that is evolving too fast to keep track of.
In a LinkedIn post, he expressed, “AI writes a ton of the small stuff, but it doesn’t know your project goals, business logic, or system design. It won’t magically know the right comparisons, conditions or algorithms.” His courses experiment with open-source large language models (LLM), GitHub Codespaces and low-barrier cloud platforms that allow students to work with AI in real environments, even if they are operating on a low budget.
Madar’s goal to make AI tools accessible is done through levelling the playing field by educating and helping students distinguish between what he calls “What is real and what is hype.” While Big Tech companies like OpenAI have a multitude of resources at their disposal, regular day-to-day people do not, unless they have a background in mathematics or computer science. This drove Madar to build a program that educates his students on what happens behind the AI curtain on which he comments, “Through a lot of research, I came up with a particular combination of technologies that I believe can indeed show my students how things work behind the scenes without a big budget.” Although unable to provide an exact rundown on the technology, he is able to provide a new angle from which to look at it.
Outside of the classroom, Madar brings coding literacy to K-12 students. In 2017, he founded a tech company designed to create educational opportunities regarding computational fluency. Like any class in school, Madar believes coding does not have to be exclusive to a specific group; it should be foundational to provide students with the option to learn skills that can be pivotal to the after-school experience of modern society’s job market. Madar’s immense bank of information led him to be an expert witness on one of Canada’s earliest software-related copyright infringement cases. That experience challenged him by forcing him to explain something technical to a non-technical audience. As he puts it, “I was analyzing code, and I tried to translate that into language that the court can understand. I’ve always loved communicating technology to a general audience.”
As AI’s growth accelerates each day, the judicial system is lacking in terms of regulating what AI can do and what data it collects. A question that is consistently on everyone’s mind: Does AI violate copyright? “That’s a can of worms, right? I do teach this in my class, but the courts are all over the place right now,” Madar explains. The legal system is particularly struggling because copyright laws were not designed with AI in mind. On top of this, the Copyright Act is not being regularly updated, meaning that fair dealing allows a good amount of leeway; limited portions of specific content are allowed to be reused, while it is not specified what ‘limited’ means.
This becomes increasingly complicated when considering that the Canadian Copyright Act focuses on “human expression” as the main form of authorship. As per the Supreme Court judgments on Intellectual Property, “an original work must be the product of an author’s exercise of skill and judgment. The exercising of skill and judgment required to produce the work must not be so trivial that it could be characterized as a purely mechanical exercise.” This begs the question: would entering a prompt in ChatGPT be considered human expression? Or, can it be copyright-stricken?
Prompts are protected under copyright laws, because they also fall under human expression. One cannot copyright strike an idea; the prompt you enter is original—hence protected—and cannot be copyright-stricken. Madar continuously emphasizes that the legal system is still working in a pre-AI world, forcing the discussion to either be this or that with no grey areas to explore. Madar expressed that the ambiguity of this is on a technical level. “After we train it, you can’t 100 per cent say that this file was in there,” he explains, “If you think of it as a recall or a remix machine, all it does is remixes, but of course, if you’re randomly remixing things, sometimes you’ll get the original back.”
In a LinkedIn article he wrote, called ‘Why coding is still essential in an AI world,’ Madar encourages collaboration. However, it is more than just working alongside AI; he discusses how it needs human observation rather than simply leaving everything up to AI. It does not replace human agency, meaning that the coders get to optimize the data presented, filter it and “ensure the results align with their goals.” He later references how web designers have already adopted and adapted to this new path, adding that designers and developers used to work in harmony. One mainly focused on visual creativity, while the other worked to “bring their designs to life.” By making AI-powered tools accessible to web designers to “transform wireframes into functional web artifacts, like HTML, CSS and JavaScript, with the click of a button” could be highly beneficial to web designers, this could also prove harmful in the sense that web designers now have to learn a new skill that was initially handled by developers: coding. Madar predicts, “As the industry evolves, designers who are proficient in both the visual and technical aspects of web design will become highly sought after, bridging the gap between creativity and functionality.”
Madar’s caution toward the uses of AI, as well as how big companies are utilizing it, is very grounded and unromanticized. AI is not just what we all see, but multilayered in how computer scientists train it. “There are actually two stages to training any AI model,” he explains, “There’s the pre-training, and then there’s the post-training. In the post-training stage, we give it the ability to converse. We give them the ability to pretend to be humans, this is where I think it’s extremely irresponsible for these big companies to actually do that.”
Despite the complexity of it all, the takeaway from Madar is that thinking and analysis matter. Whether in a class full of K-12 or university students, or even analyzing code for a court case, computational education is essential to understanding the new era that we are rapidly progressing into, with no guidance yet. Madar is passionate about technology and computer science, and his dedication shows through his work, especially in how he is trying to be an anchor that helps connect us to this new technology.