On Tuesday 21 May, HTI and the Australian Securities and Investments Commission (ASIC) hosted the ASIC x UTS: AI Regulators Symposium to discuss emerging issues and critical questions for the effective regulation of AI.
ASIC x UTS: AI Regulators Symposium
The symposium consisted of a series of discussions, including an interactive roundtable with 40 AI thought leaders in the afternoon, and culminating in the evening with a public panel led by Professor Nicholas Davis, Co-Director, HTI, featuring some of Australia’s leading regulators, including ACCC Commissioner Liza Carver, eSafety Commissioner Julie Inman Grant, Privacy Commissioner Carly Kind, and ASIC Chair Joe Longo.
Across the day, participants shared a wide range of insights, including:
While there are clear gaps in legislation and an opportunity for law reform, the effective application and enforcement of existing laws can address many of the harms arising from AI systems.
Although regulators would be better supported by law reform and additional resources, they can provide guidance, undertake enforcement action, and coordinate with other regulators on AI-related issues.
Australia can play a leadership role in the protection and enforcement of people's rights in response to harms caused by AI systems.
The regulators panel explored these issues through a discussion of their current obligations and powers, new risk areas and how AI is changing their work. The audience heard that:
Regulators are deploying AI to support staff and assist with investigative and enforcement functions.
AI is impacting regulators by creating new sources of harms and ‘turbocharging’ existing ones. Many of these issues overlap and regulators are aware of the need to work together to coordinate their responses to AI harms.
Traditional enforcement strategies are being challenged by AI for practical reasons (e.g. scale and accessibility of AI technologies; consumers unaware of rights or breaches yet burden on them to bring complaints) and legal reasons (e.g. uncertainty around liability for AI systems; application of Australian laws to foreign developers who dominant the market).
Regulators would benefit from more resources. Additional financial resources are needed for regulators to investigate and bring AI-related enforcement actions. Regulators also need to upskill staff and build technical expertise.
The first panel of the evening, Reflections on international regulatory trends from AI thought leaders, included Anna Jaffe, Director of Regulatory Affairs & Ethics, Atlassian; Lizzie O’Shea, Chair of Digital Rights Watch; and HTI’s Professor Ed Santow, facilitated by HTI’s Professor Davis.
The panel members had attended a roundtable discussion in the afternoon under the Chatham House Rule facilitated by Professor Davis. Bringing together 40 AI thought leaders from government, industry, civil society and academia, the interactive discussion focused upon questions of how existing laws apply to AI use, new and emerging AI use cases where there is a gap in laws or regulatory powers; and lessons from international regulatory approaches for Australia.
The insights shared during the roundtable and the reflections panel included that:
There is a need for law reform in key areas in response to AI, particularly the modernisation of privacy laws, as well as the importance of uplifting regulator capabilities and resources. Participants also highlight the ability of AI models to mislead and manipulate consumers as a concerning area requiring greater regulatory attention.
While gaps exist, current laws can address many harms from AI systems. However, there are significant barriers relating to enforcement. Regulation needs to address the full spectrum of AI harms with remedies for the individual consumer and at the systemic level.
Community engagement on these issues is necessary to identify AI-related problems and mitigate the worst excesses of irresponsible AI usage before they manifest at scale.
HTI has collated the key insights and take-aways from the discussions throughout the symposium into an Insight Summary: ASIC v UTS: AI Regulators Symposium.
