The Australian Ad Observatory uses machine vision and citizen-donated data to improve transparency around online advertising, helping researchers, policymakers and the public better understand how digital platforms target and influence users.
What's the problem?
Digital platforms play a major role in shaping the information people see online, yet the advertising systems that power these platforms are often difficult to observe and understand.
Researchers, governments and civil society organisations have limited visibility into the advertisements people receive, who is being targeted, and how advertising influences public debate, consumer behaviour and decision-making.
Without independent access to this information, it can be difficult to understand how automated advertising systems operate and their broader impacts on society.
What's UQ's innovation?
Professor Nicholas Carah and UQ researchers contribute to the Australian Ad Observatory, a project of the ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S) that uses research infrastructure from the Australian Internet Observatory to improve the observability of online advertising in Australia.
Australians voluntarily participate by installing an app that collects advertisements appearing in their digital feeds. A machine vision model identifies and classifies these advertisements as they appear across users' screens, creating a large-scale dataset that enables researchers to analyse how advertising operates across major digital platforms.
The project combines machine vision, digital media research and citizen participation to generate independent evidence about targeted advertising, digital platforms and automated online systems. By using AI to make digital advertising more observable, the Observatory helps improve transparency and accountability for citizens, researchers, civil society organisations and government.
What's the impact?
The Australian Ad Observatory is helping make targeted online advertising more transparent and providing researchers, policymakers and the public with new ways to understand how digital platforms operate.
The first phase of the project involved 1,909 participants and collected 328,107 unique Facebook advertisements, creating what ADM+S describes as the largest known collection of targeted advertisements encountered by Australians.
The project pioneered a new way to observe how social media advertising is targeted across populations of users, creating a unique evidence base for understanding digital advertising systems. The resulting data has enabled research into political advertising, gambling promotion, unhealthy food marketing, financial services advertising, environmental claims and online misinformation.
By providing independent evidence about how automated advertising systems function, the Observatory helps inform public debate, regulation and evidence-based policy.
Did you know?
The first phase of the Australian Ad Observatory collected more than 328,000 unique Facebook advertisements donated by 1,909 Australian participants, creating one of the largest independent datasets of targeted advertising in Australia.
Led by
Key contributors
| AI Research Strength | AI Strategy & Governance Social Impact |
|---|---|
| Industry Portfolio | Social and Economic Policy |
| Key Partners | ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S) Australian Internet Observatory Researchers and partner institutions across the ADM+S network, Centre for Digital Cultures & Societies |
| Key Publication | Angus, D., Obeid, A. K., Burgess, J., Parker, C., Andrejevic, M., Bagnara, J., Carah, N., et al. (2024). The Australian Ad Observatory Technical and Data Report. ADM+S Working Paper Series. DOI: 10.60836/bzfz-fk08 |
| Supporting Link |