Guardrails needed for AI, says Ont Human Rights & Privacy Commissions

The Ontario Human Rights Commission (OHRC) and the Information and Privacy Commissioner of Ontario are calling for the provincial government to put up restrictions on the usage of artificial intelligence in the public sector.

The use of artificial intelligence tools in government agencies across the world has skyrocketed, as has the use of AI tools among the general public, such as ChatGPT.

Many are calling for a moratorium or freeze on AI use while regulators try to come up with a way to mitigate any negative effects which may result from the use of these tools. With respect to the use of AI in government agencies, it's possible that citizens rights could be infringed if AI is allowed to run amok.

Both the OHRC & Ont. Privacy Commissioner say rules are needed for Ontario to guarantee ethical usage of AI that's supported by public trust. In their joint statement, these two commissions also pointed to the enormous benefit that AI could bring to society in areas of health, education and public safety. Nevertheless, use of AI has inherent risks if allowed to be used in a haphazard manner.

For AI to work in a lot of areas it must have access to enormous amounts of data, some or all of which could be highly personal depending on what area the AI is being used in. For example, health data. AI can  also behave in a way that could be construed as biased, or even racist, in some of the conclusions that it can arrive to when assessing data involving marginalized groups of people.

According to their joint statement, AI can also make decisions in an opaque way that's difficult to challenge. "The use of AI technologies, especially generative AI systems, may create flawed or inaccurate content that raises concerns about how government can ensure accountability for their use," reads the statement.

Across the world, regulators are attempting gain a foothold in curtailing the usage of Artificial Intelligence. In the European Union, there is a proposed Artificial Intelligence Act and in Canada the federal government has Bill C-27 and the associated Artificial Intelligence and Data Act.

But the laws contained within the federal government's bill do not cover Ontario's public sector, there is an additional need (according to these commissions) for the province to develop its own set of guardrails to constrain the use of AI by agencies falling in the jurisdiction of provincial authority.

"The IPC and OHRC appeal to the Ontario government to continue to show leadership by establishing clear and binding guardrails around the public sector’s use of AI technologies. Such guardrails must effectively address safety, privacy, accountability, transparency (including access to information), and human rights."