'While many believe AI can improve fairness, there's still work to do in building trust,' says expert
Nearly half (47%) of North Americans believe AI will eventually be less biased than human decision-making.
Canadian respondents are even more optimistic, with 52% sharing this view compared to 46% of U.S. participants, according to a survey released by AI litigation platform Alexi.
However, skepticism remains—27% of those surveyed disagree with the idea that AI will be less biased, while 24% are uncertain.
"Bias in decision-making has long been a challenge, particularly in the legal system," said Mark Doble, CEO of Alexi. "This data shows that while many believe AI can improve fairness, there’s still work to do in building trust."
Despite AI’s increasing capabilities, trust in its decision-making remains a hurdle. While 35% of respondents expect AI to handle most legal tasks by the end of 2025, there is ongoing doubt about its ability to replace human expertise.
Instead, AI is largely seen as a tool for boosting productivity, with 44% of participants anticipating that it will take over repetitive and monotonous tasks, said Alexi.
More than half (56%) of Canadian respondents reported being likely to embrace AI tools, compared to 49% of U.S. participants.
"The real opportunity for AI isn’t in replacing human expertise, but in making professionals more efficient, precise, and focused on higher-value work," said Doble. "For AI-driven companies like Alexi, the priority is not just building powerful tools, but ensuring we always work alongside humans to create better outcomes."
Overall, nearly half (47%) of Canadian employees admitted to using unregulated AI tools, found a separate survey.
At the same time, public skepticism about AI’s ability to achieve emotional intelligence remains. Only 15% of respondents consider it very likely that AI will develop emotional intelligence by the end of 2025, while 18% see this as highly unlikely, said Alexi.
With AI becoming more embedded in professional settings, the survey found strong support for AI education. A significant majority (72%) of North Americans believe AI literacy should be included in school curriculums by 2026.
Canadian respondents were particularly in favour, with 60% backing the idea, compared to 45% of their American counterparts.
The survey was carried out in February among 2,000 respondents, half in the U.S. and half in Canada.
