Concerns about “model collapse” suggest that relying solely on AI-generated data could degrade AI systems’ quality and diversity over time, potentially leading to a less effective and less diverse AI landscape, though the actual risk of catastrophic failure may be overstated if human and AI data continue to coexist.
Governments are grappling with regulating AI due to its benefits and risks. EU’s new Artificial Intelligence Act offers a pioneering, risk-based framework that other countries, including Australia, can learn from as they develop their own regulations.
Prof Mark Sanderson writes that Judge Amit Mehta’s ruling against Google acknowledges its high-quality search engine while exposing its illegal monopoly, which complicates the impact of advertising on search result quality and the potential improvements through competition and AI.
Recent research, published by Assoc Prof Timothy Graham in the journal Media International Australia, highlights how misinformation and conspiratorial narratives on social media platforms – in particular, X (formerly known as Twitter) – played a key role in the outcome of the Voice referendum.
The Australian Internet Observatory, will investigate how social media users interact and the content on their feeds. But the federal government can also help by forcing tech companies to let some light in to the closed black boxes that power their business.
Facial recognition technology is increasingly used in Australia for various purposes, with new laws expanding its use in digital identification systems. This recent survey reveals that while Australians support its use for government services and criminal identification, there is significant concern about its application for commercial purposes and a strong preference for transparency and consent.
A new study using the Australian Ad Observatory has found that unhealthy food and drinks are promoted online in ways designed to appeal to parents and carers of children, and children themselves. Additionally, young men in the study were being targeted by fast-food ads.
Meta has once again suggested it could ban links to news on Facebook and Instagram in Australia. Professor Axel Bruns discusses how a similar news ban in Cananda has produced strongly negative results for news outlets.
Developing and testing a novel suite of generative and data driven simulations, useful for depicting current and future urban scenarios, including in mobility, urban policymaking, and health domains.
This article by ADM+S Researchers Dr Zahra Stardust and co-author Prof Alan McKee shares findings from a new study title ‘Mandatory age verification for pornography access: Why it can’t and won’t ‘save the children’ published in Big Data and Society. The study looks into a common facial age estimation tool that shows such technologies are unreliable, and have a racial and gender bias.
The 2023 ADM+S Annual Report details the Centre’s achievements and progress throughout 2023 and underscores our commitment to creating knowledge and strategies necessary for responsible, ethical, and inclusive automated decision-making.
Recent research demonstrates how generative AI can be used beyond generating text. Learning more about what these tools can do will improve your digital literacy and help you understand their full impact, from benign to problematic.
This page offers basic GenAI terms and examples, developed in collaboration with ADM+S and the Office of the Victorian Information Commissioner (OVIC).
The ARC Centre of Excellence for Automated Decision-Making and Society will be hosting the 21st Chinese Internet Research Conference (CIRC 2024) in Brisbane. The theme of the event is ‘Politics and Geopolitics of Automated Decision-Making on the Global Chinese Internet’ and will be held on 17-18 June.
The Australian Ad Observatory project is asking for your help. This project evaluates to what extent online advertising targets Australian internet users.
Winners of the 2023 ADM+S Hackathon have visited Microsoft and Canva offices in Sydney to further advance their research exploring human and machine bias.
Part two of our Electronics < > Ecologies discussion on ENERGY brings engineers’ voices to the foreground, drawing on real examples of AI implementation and sustainability goal-setting in the context of Big Tech.
Information retrieval has for decades focused on finding digital documents, including documents that were born digital and those that have been digitised. But there are also enormous collections of physical documents, on paper or microfilm for example, that are not likely to be fully digitised in our lifetimes.