The generative AI fervor driven by buzzy ChatGPT has taken the tech world by storm, for its sensational capabilities and limitations alike.
But Canadian enterprises have been slow to jump on the bandwagon, unlike those in the U.S., a new KPMG survey revealed.
For this study, KPMG researched the opinions of 250 companies, 90 in Canada and 160 in the U.S, each reporting an annual gross revenue of between US$500 million and US$1 billion.
The survey revealed that only 35 per cent of Canadian organizations currently use AI in their operations, compared to 72 per cent in the U.S..
“Canadian organizations are trailing behind their U.S. counterparts when it comes to AI adoption, and it comes at a time when developments in the space are moving fast – particularly in generative AI,” said Benjie Thomas, Canadian managing partner, advisory services, at KPMG in Canada.
Over 50 per cent of Canadian respondents admit that they could be using AI more effectively. Currently only about 30 per cent are experimenting with ChatGPT or using AI in their call centres. In the U.S. it’s double that.
One of the main roadblocks that Canadian organizations face is a lack of talent. Nearly half say that they lack the in-house expertise to verify the accuracy of their AI algorithms.
Another KPMG survey, released in February, shows that over 50 per cent of employees do not trust AI at work, which can also factor in its slow adoption.
Additionally, many claim that the data sets used to train AI algorithms are either too small or too big, are missing data, are incorrect, or not properly formatted.
“The first step for any organization thinking about adopting AI is ‘get your data ready,'” said Zoe Willis, partner and national data and digital lead at KPMG in Canada.
According to Willis, this entails compiling a full inventory of the data, mapping out where it is located, assessing how accurate, recent and relevant it is, identifying current and potential data gaps, and determining who has access to that data and whether they are well-equipped to analyze it.
“Without quality data, AI algorithms are susceptible to output that is biased, incorrect, misleading, and unreliable, and the consequences for businesses include errors that lead to poor business decisions,” added Willis.
But despite over 50 per cent of Canadian respondents acknowledging the risks of making decisions based on poor quality data, only 44 per cent regularly retain independent third-party experts to assess their AI algorithms for errors and bias, compared to 75 per cent in the U.S..
A robust AI governance framework that ensures data integrity, privacy, accountability, security, and other considerations is also key to AI adoption, but only 43 per cent of Canadian organizations have one.
The need is even more critical now, as regulators globally start to clamp down on tools like ChatGPT over privacy and national security concerns.
The Canadian government, for instance, tabled Bill C-27 in June last year, in a first ever comprehensive attempt to regulate AI. The bill is yet to face Parliamentary deliberations.
Additionally, the Canadian Privacy Commissioner announced earlier this month that it will be investigating ChatGPT for possibly using personal information without permission. Last week, U.S. President Biden’s administration also called for public comments on potential accountability measures for AI systems.
The KPMG survey reveals that 72 per cent of U.S. organizations have a responsible AI framework, but as the technology evolves quickly, the need to adapt and place more guardrails is critical.
“Organizations need to have AI models that are effective, long-lasting, but also agile enough to adapt to the world around them,” said Kareem Sadek, partner, advisory, IT and emerging technology risk leader, at KPMG. “Organizations that don’t do this will be less competitive and trustworthy and will eventually fall behind.”