BACKGROUND:
Digital marketing is increasingly used to promote harmful commodities, including gambling, alcohol, and unhealthy foods and beverages, often targeting children and youth. Monitoring exposure is challenging due to behavioural targeting and highly individualised advertising experiences. To address this gap, we developed SCANNER, a deep-learning–based system that automatically detects leading unhealthy product brand logos in digital advertising. This study used SCANNER to quantify children’s and youths’ exposure to harmful online marketing in Australia.

METHODS:
Fifty-four children and youth aged 8–25 years in Australia recorded screen activity on their usual digital device for 30% of their typical screen time on one weekday and one weekend day. Screen recordings were analysed using SCANNER to quantify exposure to digital advertising for three commodity groups—unhealthy foods and beverages, alcohol, and gambling—during typical daily screen time.

RESULTS:
Children aged 8–12 years (n = 18) were exposed to an average of 22 unhealthy food advertisements, five alcohol advertisements, and four gambling advertisements for 1 hour and 35 minutes of usual screen time. Comparable exposure was observed across older age groups. Adolescents aged 13–17 years (n = 18) viewed 17 unhealthy food, three alcohol, and two gambling advertisements during approximately 2 hours of screen time. Young adults aged 18–25 years (n = 18) viewed 24 unhealthy food, seven alcohol, and seven gambling advertisements during approximately 2 hours and 37 minutes. Findings also reveal limitations of age-restricted social media bans, as around one‑third of children reported using shared devices or social media accounts belonging to older family members.

CONCLUSIONS:
SCANNER enables rapid quantification of real-world exposure to harmful digital marketing that was previously largely unseen. These findings highlight an urgent need for stronger, government-led regulation to better protect children and youth from harmful online marketing.