Shaping our Future Symposium

Bridging the AI Governance Gap

The symposium was an energetic and enlightening day of panel discussions and workshops, with the opening address from Assistant Treasurer the Hon. Stephen Jones MP underscoring the Australian Government’s support for the good use of AI as a shield, taking up the fight to fraudsters. He also highlighted the importance of regulating the activity rather than the technology when it comes to AI, the need for mandatory guardrails and a risk-based approach to regulation. ASIC Chair Joe Longo spoke about the importance of strengthening the current regulatory framework to meet the challenges of AI and the need to regularly reflect “is this enough?”

Excellent panel discussions offered important insights for organisations navigating the AI governance landscape.

The first panel with Dr. Miah Hammond-Errey, Lee Hickin, Aurélie Jacquet, and Julie Inman Grant, chaired by HTI's Prof. Ed Santow, discussed international regulatory trends and the implications for AI governance. We heard:

• Governance alone isn’t enough. We need an ‘eyes in the boat’ approach to AI use within organisations and an ‘eyes out of the boat’ focus on what is happening around leaders, such as the laws that already exist and those on the horizon.
• The AI story is a data story. Emerging technologies rely on data. For organisations to innovate responsibly, they must prioritise data governance; create a culture that asks the right questions; have the courage to create a culture of innovation; and to set a risk appetite with clear thresholds.
• Of the need for anticipatory regulators, international regulatory coherence and collaboration, and transparency (however uncomfortable) and AI standards as ‘treasure maps’ to implement and scale AI responsibly.

Next, Prof. Didar Zowghi, Bill Simpson-Young, Anna Jaffe, Chris Dolman, chaired by HTI’s Prof. Nicholas Davis discussed the essential components of AI governance. Key insights included:

• People who are using or impacted by AI systems need to be included in their design. Ethical frameworks must be created in line with product development.
• Playbooks that support changes of practices when implementing new responsible AI principles are valuable, with flexibility to respond to different use cases.
• Organisations need to take a holistic view to AI governance, it’s about the data, system, human, process and governance, while governance systems are navigating the space of top down regulation, and bottom up use cases from data.

Lastly, Dr Frances Foster-Thorpe Lizzie O’Shea Dr. Gerry Ayers, OHS & E Manager, CFMEU and Stephanie Tonkin, chaired by HTI’s Lauren Solomon explored the engagement of impacted communities in AI governance, with a pre-recorded address by Andrew Strait of the Ada Lovelace Institute. We learnt:

• While a long history of impact assessment methodology exists in a range of fields, very little consistent guidance is provided as to how this can be done well.
• Key lessons from organisational governance engagement with consumer organisations include: ensuring senior representatives are in the room; develop a long term relationship; listen sincerely; put ego to the side; and demonstrate action in response. Paying participants is important to value their work.
• We should learn from WHS settings about the obligation for organisations to consult with workers as part of hazard identification in high-risk construction work.
• Key insights from public sector engagement efforts include: a need to push beyond consultation and strive for co-governance and co-design with impacted communities; don’t talk down to participants; and assume a mutual educational process.

The AI Corporate Governance team at HTI are developing a series of resources for participants and the broader AI governance community based on the rich discussions and breakout activities. Our AI Corporate Governance Resource Library showcases many of the reports and key resources which were discussed on the day. 

We look forward to diving deeper with this community of changemakers in 2024 to make the practical changes to AI governance practise required to foster responsible AI innovation in Australia today.