Stay informed
Canadians are quick to call AI use cheating, slow to disclose their own use: poll
Most Canadians have decided that using artificial intelligence to do your work and not saying so is a kind of cheating. Getting them to admit their own use is another matter.
A new Leger survey, commissioned by the Association for Canadian Studies, the Diversity Institute at Toronto Metropolitan University and the Future Skills Centre, found 75 per cent of Canadians think AI use should be disclosed for schoolwork and 61 per cent for work. Only 32 per cent said they acknowledge their own use of AI programs.
Where Canadians draw the line is at the machine doing a lot of the work. When AI writes most of an answer, 64 per cent called it cheating, and 51 per cent said the same when it produces only part. Most (73 per cent) said using AI for ideas or editing does not count as cheating. Canadians define cheating by “authorship and effort,” Jack Jedwab, president of the Association for Canadian Studies, said in an email.
He called it “a gap between principle and practice.” Canadians value transparency, he wrote, but many “may not think their own use of AI is important enough to disclose.”
Wendy Cukier, who co-authored the report through the Diversity Institute at Toronto Metropolitan University, likened it to drivers who “judge others for speeding” while excusing their own as a necessity.
The sharpest split is by age. Among Canadians aged 18 to 24, 69 per cent use AI at least weekly, against 45 per cent overall and 37 per cent of those aged 55 to 64.
Among the youngest age cohort, 79 per cent called it cheating when AI writes most of an answer, and 54 per cent said they know someone who cheats with it, compared to only 21 per cent of Canadians overall.
Young people are also the most likely to feel hooked, with 52 per cent saying they are becoming dependent on AI, against 34 per cent of Canadians overall. Yet the youngest are the most willing to own up to their own use (45 per cent of those aged 18 to 24 and 44 per cent of people aged 25 to 34) which Jedwab wrote “suggests they’re more capable of establishing some boundaries.”
Dependence is not the same as misuse, Jedwab wrote, but it becomes a concern when someone “cedes critical thinking to AI.” Among AI users, 25 per cent said the tool already does most or all of their work.
Broad statements about responsible use, Jedwab wrote, are “insufficient.” Schools and employers need explicit rules on what help is acceptable and when it must be disclosed. In the classroom, Cukier wrote, educators may have to “rethink what we are trying to assess.”
The real issue, Jedwab wrote, is not AI itself but “the potential loss of human insight and authorship.”
The online poll surveyed 1,700 people from April 14 to 27 and the report was published this month . A margin of error cannot be associated with a non-probability sample, but Leger says a random sample that size would be accurate to within 2.5 percentage points, 19 times in 20. The Future Skills Centre, which contributed to the report, is funded by the Government of Canada’s Future Skills program.
Our website is the place for the latest breaking news, exclusive scoops, longreads and provocative commentary. Please bookmark nationalpost.com and sign up for our newsletters here.
Comments
Be the first to comment