February 07, 2024

Interoperability comparison between the proposed Artificial Intelligence and Data Act and the European Union’s draft Artificial Intelligence Act

Artificial intelligence is being incorporated into a vast array of products and services that are offered and used in multiple jurisdictions around the world. It is therefore a critical public policy objective to ensure that Canadian domestic statutory frameworks and other instruments designed to ensure the responsible use of AI align and are interoperable with regulatory frameworks adopted by major trading partners.

International standards for regulating artificial intelligence are in their formative stage. In Canada, Bill C-27 sets out the federal government’s proposed Artificial Intelligence and Data Act (AIDA), which is currently the subject of testimony before a parliamentary committee. As part of his testimony, the Minister of Innovation has proposed substantive amendments to AIDA through a letter to the INDU committee on November 28, 2023.

In the EU, the full text of the proposed Artificial Intelligence Act (AI Act) has recently been made public, so it is now possible for Parliamentarians, businesses, civil society and all other stakeholders in Canada to assess to what extent AIDA is aligned and interoperable with the AI Act.

To facilitate this interoperability assessment, AccessPrivacy has created a table that sets out a non-exhaustive list of requirements under the EU AI Act and AIDA, a summary of the key distinctions in such requirements, the interoperability impact, and examples of such impact.

The table highlights that, if enacted in its current form, AIDA will impose material regulatory obligations on a substantially broader range of artificial intelligence systems and machine learning models than under the AI Act. More broadly, the table highlights that the AI Act represents a more targeted, proportional, and risk-based approach to regulating AI compared to the approach reflected in AIDA.