Responsible AI

The responsible use of AI is fundamentally about defining basic principles, managing their use and putting them into practice. The goal is to ensure the outcomes of AI initiatives and solutions are safe, reliable and ethical.

AI’s widespread accessibility marks a major opportunity, but also introduces challenges. The lack of transparency in data sources and outputs, and a largely unregulated landscape to date, creates potential risks for organizations. These risks range from privacy and copyright infringements and security threats to reputational and financial consequences.

An AI governance framework that balances positive impacts with risk mitigation is critical to unlocking AI's full potential.

CGI defines three pillars of responsible AI as follows:

• Robust: Ensure the reliability and safety of outcomes and compliance with applicable laws and regulations by enabling rigorous design, development, testing and operations processes supported by privacy and security policies and practices.
• Ethical: Ensure the goals of AI use are aligned with human values and that actively foster fairness and inclusiveness, and that the results sustainably benefit individuals, society and the environment.
• Trustworthy: Ensure AI use is based on explainable and interpretable models, as well as transparent input and outcomes. This includes taking responsibility for the AI technologies an organization controls.