From promising messages to real-world outcomes: Unclouding the Future enters its trial phase

NCYSUR research fellow Dr Giang Vu reports on progress towards a vaping education RCT

Vaping education messages can attract attention and receive positive feedback, but does that mean they influence what young people know, think or do?

Unclouding the Future is entering a new phase to investigate this question. Following the development and initial evaluation of AI-assisted, youth co-designed vaping education materials, the research team is preparing to conduct a randomised controlled trial involving approximately 900 young Australians aged 16–25.

Combining generative AI with youth involvement

In the project’s earlier phase, generative AI was used to help produce initial text and visual concepts for vaping education materials. Young people then reviewed and reshaped the materials through interviews and focus groups.

Their input influenced not only the wording and appearance of the messages, but also their tone, focus and relevance. The process highlighted an important distinction: generative AI can help researchers develop ideas efficiently, but young people bring lived perspectives and contextual understanding that technology cannot be assumed to reproduce.

Researchers retained responsibility for checking the accuracy and safety of the information and making final decisions about the materials.

An initial evaluation found that young people rated the AI-assisted, youth co-designed materials as having greater perceived message effectiveness than existing vaping education materials. This was a promising result, but perceived effectiveness reflects people’s immediate reactions to a message. It does not establish whether seeing the materials changes vaping behaviour or produces lasting improvements in knowledge and attitudes.

Testing what happens over time

The new trial will move beyond immediate message ratings by following participants for 12 months.

Participants will be randomly allocated to receive one of three types of content: materials developed with direct input from young people; materials informed by AI-generated simulated participants; or control materials unrelated to vaping. This comparison will help the research team examine what direct youth involvement contributes and whether emerging AI-based approaches can usefully support formative research with real people.

Participants will receive materials directly through MMS/SMS and/or WhatsApp over either eight or 16 weeks. They will complete surveys at baseline and again after one week, two months, four months and 12 months.

The study will examine whether the materials influence outcomes including vaping behaviour, attitudes towards vaping and knowledge about vaping. It will also consider outcomes such as intentions, susceptibility, perceived harms and social norms.

Comparing eight and 16 weeks of delivery will address another practical question for health communication: does a longer period of repeated exposure produce different outcomes from a shorter campaign? The researchers hypothesise that sustained exposure may strengthen the effects of the materials, but the trial is designed to test rather than assume this.

Informing future health communication

The trial could provide evidence on several questions relevant to researchers, health services and campaign developers.

First, it will examine whether materials that young people perceive as persuasive can influence behavioural and other outcomes over time. Second, it will investigate whether campaign duration matters. Finally, it will provide evidence about how generative AI might be used responsibly in public health communication and where direct involvement from the intended audience remains essential.

The project is especially interested in the value of combining technological tools with meaningful human involvement. Youth co-design is not simply a final check on materials that researchers have already produced. When undertaken meaningfully, it can change the assumptions, language and creative direction underpinning the work.

By comparing different development approaches within a longitudinal trial, Unclouding the Future aims to move the discussion from whether AI can produce health content to the more useful questions of how it should be used, how its outputs should be evaluated and whose perspectives need to guide the process.

Acknowledgements

Unclouding the Future is funded by the Australian Government through the Medical Research Future Fund’s Preventive and Public Health Research Initiative, under the 2023 Maternal Health and Healthy Lifestyles grant opportunity.

The research team also acknowledges the young people who contributed their time and perspectives to the development and refinement of the campaign materials.

For further information about the project, contact the research team at g.vu@uq.edu.au.

In this issue