Dr Maria Milosavljevic GAICD FACS

The UTS Human Technology Institute was honoured to host the AI Policy Roundtable this week. This extraordinary gathering spanned four continents, and included some of the world's leading and most influential AI policy experts. I've been buzzing from this brilliant discussion ever since... Here are a few key insights I've been reflecting on: 🤜 The last 6-7 years has seen a lot of reform on AI law and policy. The vast majority of that reform is here to stay. We should spend less time talking about what the rules for AI should be, and more time ensuring those rules are understood, applied and enforced. 🤜 Australians are among the 80% of the world's population who don't live in the US or China... and so we should work more collaboratively with countries that face similar challenges and opportunities on AI that we do. 🤜 When reform stalls at the national or federal level, this usually means that the states (ie, sub-national governments) pick up the slack. The observation that struck me most forcefully was this: it's often said that AI law reform isn't a vote swinger... this is both true and misleading, because voters are unforgiving when a government fails to address known risks and threats. Time and again, Australians have made clear that they want a balanced approach to AI that takes opportunities, and addresses risks. Governments should listen. This was a discussion under the Chatham House rule, so I wish I could attribute wisdom to the clever people who made these observations! Instead, I'll make do by giving special thanks to the people who travelled the furthest to join this discussion: Gaia Marcus FRSA, R. David Edelman, Marc Rotenberg, Ambassador Philip Thigo, MBS & Daniel Stone. Brilliant contributions were made also by Lyria Bennett Moses, Valeska Bloch, Simon Burns, Sam Burrett, Elizabeth Clarke of KPMG, Bryony Evans, Chandni Gupta, Dr Jane Hogan, Aurélie Jacquet, Anna Jaffe, John Livingstone, Georgia Lovell, Isabella Newell, Rob Nicholls, Anna Nitschke, Nandita Sharma, Bill Simpson-Young, Ramona Vijeyarasa, Melisande Waterford, Johanna Weaver, Simon Writer & Evana Wright. I'm so fortunate to have such brilliant colleagues to bring this event together, especially Sophie Farthing, Sarah Sacher, John Croker, Amy Pham & Lauren Perry. Finally, but most importantly, I join with HTI co-founder Nicholas Davis in thanking the Minderoo Foundation, skilfully represented by Tim Rutherford & Michelle Park, for making this event possible.

250

16 Comments

Trust, Attitudes and Use of Artificial Intelligence | prepared by KPMG in collaboration with The University of Melbourne This report examines the evolving landscape of #AI adoption, #publictrust, and organizational deployment strategies, highlighting the nuanced interplay between technological capability, regulatory frameworks, and societal perception. Drawing on survey data from over 5,400 participants across #Australia, #theUnitedKingdom, and #theUnitedStates, and interviews with 87 senior executives, the study situates AI not merely as a tool but as a socio-technical system where trust, governance, and ethical alignment directly impact adoption rates and value creation. The research frames AI deployment within three interdependent dimensions: organizational readiness, user perception, and accountability mechanisms, revealing systemic disparities across sectors and geographies that influence both uptake and efficacy. Quantitative findings underscore the critical role of trust in AI adoption: 64% of respondents expressed cautious optimism about AI’s benefits, yet only 38% indicated strong confidence in existing oversight mechanisms. In organizations actively deploying AI, 71% reported integrating ethical review boards or AI governance committees, yet only 29% had formalized explainability protocols for end-users. Cross-sector analysis highlights that highly regulated industries such as finance and healthcare demonstrate higher trust scores—averaging 58%—relative to #consumertech and #retail, where confidence falls to 41%. Moreover, AI literacy emerged as a significant predictor of acceptance: participants with moderate to high AI familiarity were 2.3 times more likely to endorse AI-assisted decision-making in critical operations, emphasizing the interaction between competence, transparency, and trust. The report introduces a conceptual framework of “Trust-Enabled AI,” integrating perceptions, governance structures, and adoption metrics into a single evaluative model. Pilot case studies illustrate that firms embedding continuous monitoring, algorithmic transparency, and stakeholder engagement in AI systems achieve adoption rates 22–27% higher than peers with minimal governance interventions. Additionally, public sentiment analysis reveals that narrative framing and proactive communication about AI limitations can improve perceived legitimacy, thereby mitigating reputational and #operationalrisks. In summary, AI adoption and efficacy are inextricably linked to the triad of trust, competence, and governance. Building resilient AI ecosystems requires organizations to implement robust oversight, enhance user literacy, and align algorithmic decision-making with societal values. The findings signal that without deliberate #investment in trust infrastructure, AI #risks engendering skepticism and uneven uptake, undermining its transformative potential across both public and private sectors.

41

1 Comment