Background Patient and public involvement (PPI) has become embedded in health research. Although evaluation should be an integral part of PPI, these are often reported as descriptive accounts. Utilizing creative approaches and embedding reflexivity along the way, we undertook a participatory evaluation of a co-production process with public contributors to gain insights into a project aimed at improving public understanding of artificial intelligence (AI) and data science. The public has limited awareness of how these technologies work. To improve public understanding, the research team at Imperial College London, consisting of researchers and health professionals alongside 30 public contributors, worked together on a PPI project to co-design and co-develop an information resource.

Methods We used creative approaches for the participatory evaluation and embedded reflexivity along the way for a potentially more ethical and accessible involvement of public contributors. Firstly, we utilised the photovoice. For three months, public contributors and team members captured photos alongside a description of how this photo presented their experience. A professional photographer supported this data collection. Secondly, using over 60 included photos (alongside paintings and poetry), we met to discuss and analyse the experience of being involved in the project. The professional artist captured this discussion together in the mandala- a geometric configuration of symbols representing the experiences, from the outside to the inner core, through different layers.

Results Everyone’s experiences were spread into six interconnected layers on the mandala and embodied as a connected journey. The chosen title, ‘Our Journey into AI.’, reflected the feeling of ownership among public contributors. Layers number: 1) (from inside) represents public contributors’ opportunity to learn new things and develop their confidence in data science and AI;2) shows interaction with others and working as a team; 3) is about working digitally throughout the project. From layer four onwards, the issues broaden up and start including those that underpin public contributors’ involvement.4) is about why they got involved in the project; 5) shows that they expect an inclusive and accessible involvement process. The last layer is about the ethics of data science and AI that researchers should follow.

Conclusion Through utilizing creative approaches and reflexivity as part of our participatory evaluation, we found that the co-production with public contributors around the complex and often abstract topics like data science and AI can be conducted in a meaningful way to lay members of the public. Co-production process ensured a sense of co-ownership.