PROJECT SUMMARY
Generative Authenticity
Focus Areas: News & Media, Mobilities, Social Services, Health
Status: Active
Authenticity is a key problem for understanding and managing the impacts of generative AI and synthetic media in society, and a central target for automated decision-making systems in the information and media environment. From trustworthy news reporting to identity verification for social services and the everyday risk of scams, generative AI and synthetic media present significant real-world implications for practitioners, institutions, and publics in Australia and elsewhere.
A wide range of technical solutions collectively understood as authenticity infrastructure promise to address these issues; but if adopted and embedded at scale, some of these solutions could have potentially significant downstream effects on stakeholders and implications for society.
This project will critically examine the assumptions underpinning these developments and debates, assess the technical and legal challenges associated with them, and explore novel technical responses that contribute to more responsible, ethical and inclusive ADM systems.
In doing so, the project will draw on the multidisciplinary expertise of the Centre and our partners explore authenticity as both a socio-technical challenge and as a contested cultural idea. We address these challenges in practical and experimental ways within the innovative and Generative AI Test Range environment. It will also examine what happens after any determination of authenticity, including mechanisms for explaining and communicating determinations and increasing trust in such measures.
PROJECT OBJECTIVES
- Produce a cross-disciplinary understanding of the problem of authenticity in the context of Generative AI;
- Study and map the field of Authenticity-as-a-Service (AaaS), providing a detailed account of its infrastructure, operations, and political economy;
- Analyse how the integration of authenticity infrastructure is already playing out in practice in specific sectors, and impacting or likely to impact specific communities;
- Within the Generative AI Test Range environment, simulate and evaluate competing ADM techniques for addressing the problem of authenticity in a range of real-world scenarios; and
- Build on our findings to develop improved tools and techniques, and produce and share guidelines for explanation and communication for a range of stakeholders and audiences.
MORE INFORMATION
RESEARCHERS
Prof Christopher Leckie
Project Co-Leader and Chief Investigator,
University of Melbourne
PARTNERS
Australian
Broadcasting
Corporation
