On 14 June, Jun gave a keynote talk at the Oxford Computer Science Conference 2022 about our 5-year journey of the development of age-appropriate AI for children.
Our journey started at the end of 2016 when AI for children has been hardly discussed at that time. Our journey started with our discovery of the extensive data tracking associated with mobile apps used by children, which drove us to start our very first use study with children and the publication of "`I make up a silly name’: Understanding Children's Perception of Privacy Risks Online".
This pioneering study looking at the perception of data privacy by children aged 6-10yo gave us crucial insights regarding children's ability to recognise and cope with online data privacy risks and their knowledge gaps. At the same time of this investigation, children were increasingly being exposed to greater online privacy risks, where they are being constantly monitored and exposed to algorithmically-manipulated online content, or decisions made about their opportunities.
In 2018, we broadened our research scope, from data privacy to algorithmic risks for children, and made a range of critical scoping studies, including
- identifying how best to design tools to support their risk coping
- assessing how algorithmic systems are affecting children's digital experiences and identifying 5 key principles for designing age-appropriate AI systems for children
- examing how to support good designs
Jun concluded the talk by discussing the critical paradigm shift that is needed and being facilitated by the recently funded Oxford Martin School Ethical Web and Data Architectures (EWADA) project.
It was really exciting to be back in person and talk about the human-centred research approach with Oxford CS students! Thanks for the invite from the committee!