A number of research projects are currently using AIO for data collection and analysis. Topics include advertising, elections, gambling, recommender systems and more.
The Australian Ad Observatory
The Australian Ad Observatory, a project of the ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S), uses the Mobile RoverAIO’s Mobile Observatory Ad Toolkit screen-tracking tool and the Browser Rover browser extension to collect the advertisements that participants seeencounter on social media platforms. More than 2,000 volunteer Australian Facebook users have contributed over 700,000 ad observations. The project examines political advertising during Australian elections, gambling and alcohol advertising targeting, and the content of advertisements on Facebook, Instagram, and TikTok. An alcohol advertising study, based on 39,820 ads from 1,123 advertisers collected over 1 year, demonstrated the capacity for sustained citizen-science data collection at scale. By tracking advertisements across platforms, researchers have documented how targeting mechanisms operate differently across services and how individual users’ advertising experiences diverge based on their demographic profiles and platform histories. The project demonstrates how data donation can produce research that is impossible through platform-centric methods alone, since the advertisements it studies are “dark”, i.e., visible only to their intended targets and absent from any public ad library.
The For You
The “For You” is an ARC Discovery project led by researchers at Queensland University of Technology (QUT) and the University of Sydney. The project investigates how TikTok’s recommendation algorithm influences Australian users and shapes digital culture.
The study is organised into three research streams:
- Platform Analysis: examining the technical mechanisms behind TikTok’s content recommendation system and how they shape the information environment.
- User Experiences: studying how individuals interact with and are influenced by algorithmically curated content on TikTok.
- Creator Strategies: investigating how content creators adapt their practices in response to platform algorithms.
The project combines user-centric data donation methods with platform-centric analysis, using data download packages (DDPs) alongside screen tracking to compare what content users post and seek with what TikTok’s algorithm actually serves them. This mixed-methods design addresses a question that platform-centric methods alone cannot answer: how does the algorithmically curated experience differ from what users actively seek or create?
Alcohol and Gambling Advertising on Facebook
A 2024 study by researchers at the University of Queensland, FARE, the Alliance for Gambling Reform, and VicHealth used theMobile Rover to investigate how alcohol and gambling companies target people most at risk of harm on Facebook. The study piloted a novel digital data donation method in which ten participants at risk of alcohol or gambling harm installed the app on their devices to collect the advertisements they saw.
The findings revealed that Facebook tagged participants with up to 25 alcohol-related and 41 gambling-related advertising interests, even when those individuals were actively trying to reduce their consumption. A total of 201 alcohol companies and 63 gambling companies uploaded targeting data on the ten participants. For one participant trying to reduce high-risk alcohol use, a quarter of all ads captured were for alcohol; for another experiencing gambling harm, almost two-thirds of captured ads were for gambling, sometimes 15 or more in a row.
This study demonstrates both the power of user-centric data donation methods and the Mobile Rover’s ability to capture “dark” advertisements (those visible only to their intended targets and absent from any public ad library). The research directly contributed to public policy advocacy, illustrating how AIO tools can support not only academic research but also evidence-based public health interventions.
Political Advertising Monitoring
During the 2025 Australian federal election, researchers from the Australian Ad Observatory tracked the often invisible world of digital political advertising across Facebook, Instagram, and TikTok. Using the Mobile Rover screen-tracking tool and the Browser Rover extension, a panel of ordinary Australians agreed to download an ad-tracking app to record the ads they encountered. In parallel, the team used the PoliDashboard tool (an open-source application that aggregates transparency data from Meta) to identify patterns in political ad spending, message type, tone, and reach.
The research uncovered a striking prevalence of third-party political advertising: for every ad from a registered political party, there were roughly 1 from a third-party entity. Among the most concerning trends was a rise in astroturfing – campaigns designed to mask corporate or political sponsors behind names that suggest grassroots citizen movements, such as “Australians for Natural Gas” (found to be working with the Coalition’s internal pollster) and “Australians for Prosperity” (backed by wealthy donors and coal interests).
Fewer than 2% of the ads participants saw were relevant to political topics, showing the highly targeted nature of modern digital campaigns. This case study demonstrates how the AIO’s combination of user-centric data donation, screen tracking, and platform transparency tools can reveal advertising dynamics that are invisible to any single method alone. The findings also underscored the need for meaningful observability and truth-in-political-advertising laws in the digital age.
