The Effective and Ethical Development of Artificial Intelligence: An Opportunity to Improve Our Wellbeing: Appeal Algorithmic Decisions

We thank the Australian Council of Learned Academies for this opportunity to submit a policy input paper for consideration in your preparation of the Horizon Scanning Report on AI for the Australian Commonwealth Science Council. We have recently conducted a research project examining regulatory approaches to Artificial Intelligence (‘AI’). Our most recent output from this project is, ‘Nudging Robots: Innovative Solutions to Regulate Artificial Intelligence’ (2017) 20(2) Vanderbilt Journal of Entertainment & Technology Law 385 (attached). Our research has considered the involvement of AI in legal decision making and concluded that this was an area that should be regulated. Our paper outlines rather extensively a range of options for regulators and proposes a risk-based approach to regulation. That is, public regulators should be alert to the spectrum of risks posed by specific applications of AI and adopt targeted strategies in their regulatory approach in order to address the risks identified. This submission outlines the risks arising with automated legal decision making, proposes that human involvement in legal decision making processes is vital to address the limits of algorithmic justice, and proposes consideration of a model similar to that in Europe. We support the development of regulation allowing challenge of algorithmic decision making by AI.