AI / Machine Learning Algorithms: Bridging the Gap between Policy and Practice

This Capstone project surveyed the rapidly changing landscape of artificial intelligence (AI) to assess how companies and governments address AI-related risk and promote ethics in their AI development or use. The methodology included expert interviews, targeted research on private sector companies, and an analysis of published AI risk management frameworks. The landscape analysis revealed four key findings:

1. AI is the subject of significant attention, but that heightened attention doesn't necessarily bring with it a more thorough understanding of the AI technology itself;
2. Awareness of nuanced AI risk only comes from hands-on, technical experience;
3. There is a gap between companies' ethics practices and the many published frameworks, and
4. The gap between policy and practice is difficult to analyze - a company's actions may differ from their public messaging, making assessing a company’s commitment to AI ethics very difficult.

Therefore, the Capstone team first recommended governments educate the public on AI risks, like misinformation and privacy, and require schools to teach AI fundamentals. Second, public-private partnerships and a cross-sector information sharing hub for AI best practices should be established. Third, AI ethics evaluation is hard to generalize, so regulators should allow for more tailored evaluations. Fourth, governments should build robust auditing practices to disincentivize unethical practices.