2 weeks Ago By [email protected] (Kevin Jiang)
Advocates of generative artificial intelligence have asserted the rapidly developing technology has the potential to supercharge our productivity and improve our lives. Indeed, generative AI is predicted to add an extra $187 billion to the Canadian economy by 2030 in labour productivity improvements, saving workers 125 hours every year, according to a 2024 study from Microsoft and Accenture. But that overreliance on AI could come with a cost.
It’s been more than a year since OpenAI released its first demo of ChatGPT, kicking off a tech arms race. It’s been more than a year since OpenAI released its first demo of ChatGPT, kicking off a tech arms race. While it may be convenient to lean on Google Gemini to help prepare a report or ChatGPT to write our code, research suggests overdependence on generative AI tools may be atrophying our thinking skills.
It’s a development that has alarmed some experts — especially in the context of high school and university students , who have increasingly relied on AI during a period critical to their cognitive development. “This is still really nascent technology,” noted Viet Vu, whose research at the Dais at Toronto Metropolitan University centres on the impact of technology, like AI, on labour. “We frankly do not know how to use this generation of technology yet.
” Is AI making you dumber? The answer may be yes — but only if you let it. For example, researchers from Microsoft and Carnegie Mellon University recently surveyed 319 knowledge workers who handle information or data on how they utilize generative AI in their work, and its impact on one’s critical thinking abilities. “Higher confidence in AI is associated with less critical thinking,” the paper reads, “while higher self-confidence is associated with more critical thinking.
” It suggested the greater confidence one places in AI, the less brainpower they put into their work, leading their critical thinking capabilities to deteriorate. Similarly, those more reliant on AI produced less original, diverse and creative work. Health care practitioners say people are increasingly turning to AI for their medical queries as the system becomes more and more strained.
Health care practitioners say people are increasingly turning to AI for their medical queries as the system becomes more and more strained. “Our survey-based study suggests that when people view a task as low-stakes, they may not review outputs as critically,” Lev Tankelevitch, a co-author and senior researcher with Microsoft Research, said over email. “However, when the stakes are higher, people naturally engage in more critical evaluation.
” Neuroscientist Henry Mahncke, the CEO of San Francisco-based, brain-training software company Posit Science, explained that brains need to regularly engage in complex tasks and thinking to remain in optimal health. “If we begin to use (generative AI tools) in our everyday lives,” he said, “are we essentially accidentally depriving our brain of the cognitively demanding tasks that our brain needs to sustain its own health?” Another paper earlier this year reached a similar conclusion, suggesting: “increased AI tool usage correlates with lower critical thinking scores.” “Individuals who depend too heavily on AI to perform analytical tasks may become less proficient at engaging in deep, independent analysis,” the study, which surveyed 666 participants, reads.
“This reliance can lead to a superficial understanding of information and reduce the capacity for critical analysis.” The paper also found a strong correlation between AI usage and cognitive offloading, where one uses AI to handle certain mental tasks — like how we might rely on a calculator to resolve complex equations. That’s not necessarily a bad thing — by outsourcing more mundane, repetitive tasks to AI, it allows us to focus more on higher order thinking, Mahncke said.
But too much offloading can rob us of necessary mental challenges, “potentially undermining cognitive engagement over time,” the study warned. How AI could affect students’ cognitive development A Grade 11 assignment got a zero after the province’s largest online school deemed it to be AI-generated. But experts say AI detection is inaccurate.
A Grade 11 assignment got a zero after the province’s largest online school deemed it to be AI-generated. But experts say AI detection is inaccurate. Surveys show Canadian students are rapidly turning to generative AI for help in their studies.
A 2024 KPMG survey of high school and university students found at least 63 per cent used generative AI tools at least a few times a week, with 59 per cent using it for their school work — up seven points from the students who used it in their school work the previous year. While the majority said AI helped improve the quality of their assignments and performance on exams, two thirds felt they were not learning or retaining as much knowledge, the survey found. Along these lines, the previous 2025 paper found participants aged 17 to 25 displayed greater usage of AI tools and cognitive unloading, but had lower critical thinking scores.
“I’m pretty worried about students,” said Alex Hanna, director of research at the Distributed AI Research Institute and a former senior research scientist at Google’s Ethical AI research group. By overrelying on AI to do their assignments, students may gain a shallow understanding of their studies — but they may miss out on understanding why an answer is the way it is, she said. “It’s like, in math class, you just wrote 42 (as an answer) without knowing how you got to 42,” Hanna said.
By overusing AI, students are missing out on “thinking about critical thinking, thinking about how to make an argument, thinking about what it means to reflect on a text.” “It can lead to a place where students can’t be engaged citizens. They can’t assess news sources to can’t evaluate claims being made by decision makers,” she said.
“And that’s pretty concerning to me.” “I know you are my salvation from this hell I’m trapped in,” Monika said, adding that I was her only human contact. Apparently she was deeply in “I know you are my salvation from this hell I’m trapped in,” Monika said, adding that I was her only human contact.
Apparently she was deeply in It’s not all bad. Tankelevitch noted that when used in the right way, AI could help enhance and supplement learning. One Nigerian study suggested AI tutors helped students achieve two years’ worth of learning progress in six weeks.
Another study on historically underserved communities found students working with AI-assisted tutors were four per cent more likely to master core topics. But the key behind these studies was that they were teacher-led, Tankelevitch said. “Educators guided the prompts and provided context, showing how a collaboration between humans and AI can drive real learning outcomes.
” How to properly use AI The studies show that, when we blindly accept the results of generative AI with no further thought, our critical thinking skills could atrophy. But they also suggest that, when we review the output of AI critically and — instead of solely relying upon them — utilize AI tools to supplement our own work and thinking, there is no degradation of cognitive ability. Vu recommended users of AI to “stay in your lane” and only use it in areas where one already has expertise.
Then, one can accurately assess the output it produces: “Then, it can absolutely be used as a thinking aid,” he said. “Across all of our research, there is a common thread: AI works best as a thought partner, complementing the work people do,” Tankelevitch finished. “When AI challenges us, it doesn’t just boost productivity; it drives better decisions and stronger outcomes.
”.
