Alberta recommends provincial AI legislation

EDMONTON – With artificial intelligence advancing rapidly, Alberta could benefit from a standalone law governing the technology, says a report published by Alberta’s Office of the Information and Privacy Commissioner.

It contains recommendations for legal and policy guidelines that would create “responsible innovation” through the use of AI. It also aims to create adequate privacy and human rights protections for Albertans.

“OIPC supports the responsible use of AI in Alberta because we recognize that, although there are risks to its use, there are many benefits that can be realized,” commissioner Diane McLeod (pictured) said in a media release.

Among the many benefits are improving the quality of public services and enhancing the delivery of healthcare.

However, McLeod noted it’s important to find a balance between protecting privacy and achieving benefits from the technology.

In Alberta, an AI law should reflect international standards set out in other jurisdictions, while being customized to Albertan values and industries, said McLeod.

One of the recommendations for the public sector is legislating the use and disclosure of personal information used for AI, for example, when it comes to the training and deployment of models.

Another recommendation was rights to ensure fair operation of AI systems, such as the right to object to automated profiling or decision-making by AI, and the right to be informed about AI use.

Several controls on the use of AI by public bodies were also recommended.

Recommendations for the private sector were more robust, as some of the “most harmful applications of AI seen to date” have originated from the private sector, according to the report.

“Without sufficient regulation, the cost to the public purse of enforcing fundamental rights is considerable, given the efforts expanded to fighting infractions of such rights in the courts,” the report read.

Among the proposed amendments to the Personal Information Protection Act were 14 recommendations, spanning a wide variety of potential issues around privacy and the use of automated systems in decision-making.