Likely Stories: AHRC funded studentship
Can we predict the persuasiveness of online media based on composite elements such as such as musical arrangements, storylines and imagery? The Likely Stories project explores the use of AI in detecting media bias and predicting its effect on marginalised young people in the UK and Pakistan, in partnership with the National Centre for AI in Islamabad. The project responds to a lack of research into using machine learning to understand the aesthetics of persuasion in audio-visual narratives in the short form videos common on social media, especially in relation to marginalised audiences. How can arts and humanities theory inform and enhance machine learning approaches to problems of media bias and persuasive intent?
Likely Stories PhD studentship is funded by the North West Consortium Doctoral Training Partnership and the Arts and Humanities Research Council.
Unhealthy bias: Wellcome trust seed fund project
Research into public health campaign impact (Gregory et al., 2010) suggests that failures in persuading UK South Asian heritage families contribute to health problems, such as heart disease. Algorithmic responses to problems of media bias have the potential to challenge biases. However, for these data-led methods to be effective requires proper examination of the ‘normative technocultural beliefs’ that obscure ‘raced and gendered’ influences on online behaviour (Brock, 2015:1087).This project will use media theory and its methods to inform machine learning (ML) approaches to NHS public health online video efficacy for South Asian audiences.
Unhealthy Bias has been awarded seed funding by the The Northern Network for Medical Humanities Research and the Wellcome Trust.