Critics question Canada’s approach to AI voluntary code of conduct consultation
The critics highlight concerns about the need for more transparency, a rushed process, and inadequate public engagement in the government’s generative AI consultation.
The Canadian government has been criticised for its consultation process on a voluntary code of conduct for generative AI systems for being too rushed and not transparent enough. The ongoing consultation has been characterised as both unusual and veiled, with critics arguing that it lacks transparency, inclusive democratic engagement, and sufficient public awareness, casting uncertainty on the government’s dedication to the responsible advancement and implementation of AI.
There are additional worries that the government is not dedicating enough time to consult with experts and the public regarding this vital matter. Concerns have emerged regarding the forthcoming AI legislation, criticised for its outdated nature and limited input from various stakeholders. Some assert that a thorough overhaul of the legislation is required to clearly define the AI technologies it will oversee and the regulation methods. The legislation’s introduction prior to the emergence of generative AI systems like ChatGPT has added weight to these criticisms.
About the Canadian voluntary code of conduct
The voluntary code of conduct aims to establish trust in AI systems and pave the way for a seamless transition into compliance with upcoming regulatory rules. Some key provisions of this code include the necessity for AI developers to make AI-generated content distinguishable from human-created content and the requirement for human oversight to ensure responsible AI system usage, aligning with the European Union’s call for online platforms to label AI-generated content in the fight against misinformation. Additionally, the code mandates companies to prevent malicious or harmful utilisation of their AI systems, address bias risks, and guarantee that AI systems are easily distinguishable from humans. This code will be implemented before the enactment of Bill C-27, which includes the Artificial Intelligence and Data Act (AIDA), focusing on privacy legislation with AI components.