Analysis of Artificial Intelligence and Data Act Based on Ethical Frameworks

Title Analysis of Artificial Intelligence and Data Act Based on Ethical Frameworks
Author Sun Gyoo Kang
Affiliation National Bank of Canada
Region/Country Canada
Pages 1-28
Download
Permalink https://www.ijlet.org/4.4.1
DOI
Keywords Artificial Intelligence; Artificial Intelligence and Data Act; Risk-based Approach; Canada; EU
Abstract Canada has recently introduced the Digital Charter Implementation Act of 2022. Both at the federal level as well as at the provincial level, different governments are trying to move forward with emerging technology by amending and implementing laws and regulations to put a safety net against various emerging risks, but at the same time, promote the growth of the innovative industry in Canada. Indeed, as one of the first countries, Canada introduced legislation regarding artificial intelligence (Artificial Intelligence and Data Act), and the goal of this paper is to check if the new legislation would encompass the basic artificial intelligence ethical frameworks such as privacy, accountability, transparency/explainability, fairness, and safety & security, which were recommended in three (3) different ethical declarations on artificial intelligence, among many others. The Artificial Intelligence and Data Act contains important sections incorporating accountability, transparency/explainability, fairness, and safety & security. Considering a risk-based approach targets high-impact systems and requires risk management and monitoring against the risk of harm and biased output. It also requires persons responsible for high-impact systems to communicate all the necessary information publicly and to be audited in case of concern with its artificial intelligence systems. The new legislation imposes administrative monetary penalties and offenses similar to the proposed regulations in the EU. As for privacy, similar to the EU, it is outside of the new law’s scope, but another law covers the topic independently already Nevertheless, the Artificial Intelligence and Data Act also has some pitfalls. It lacks clarity and specific requirements found in the law of the EU. Furthermore, the scope is an issue as it only covers private sector actors, and there is doubt about the real independence and neutrality of the commissioner.