The ethical challenges of using Artificial Intelligence to understand social media habits and preferences: the examples of Libya and Cambodia

Summary:

BBC Media Action is engaging with artificial intelligence (AI) in our research efforts to handle the volume and complexity of social media data in two pilot countries, Cambodia and Libya, in view of designing stronger social and behavioural change media programming. Using AI in research presents ethical dilemmas. This presentation will highlight these ethical issues and present BBC Media Action's efforts to design a framework that guides its use of AI in research.

Background/Objectives:

BBC Media Action has collaborated with Loughborough University to develop ethical guidelines to be used throughout the research process to evaluate the ethical limitations of a study. These were adapted from ethics principles combined by Chris Alen Sula (2016), and suggest a framework that identifies seven ethics principles to guide research involving AI. The framework is a practical guide that provides standards to researchers to use throughout the research process.

Description Of Intervention And/or Methods/Design:

BBC Media Action is increasingly using social media content to engage and reach audiences. Our projects in Libya and Cambodia have almost exclusively used this medium to do this. To understand audiences, it has been challenging to use traditional research methods - online surveys with users have low response rate, and research on the ground is challenging in countries like Libya due to the volatile security situation. Therefore, BBC Media Action has been exploring using AI in research to obtain in depth information about the audiences we are looking to reach through understanding the information ecosystems that people inhabit online, and how our content manifests within those, in view of designing stronger social and behavioural change media programming. However, using AI in research presents ethical dilemmas. This presentation examines BBC Media Action's efforts to design an ethical framework to guide its use of AI in research.

Results/Lessons Learned:

Case study - In Libya, BBC Media Action used AI in order to refine the editorial proposition for our social media platform, El Kul, to make it more appealing and engaging to young Libyan audiences. AI has allowed for a better understanding of the content that target audiences are consuming across different social media platforms as well as drivers to engagement by different aspects (e.g. tone, type of content, language) including for news content. This has provided a depth and scale of insights into audience preferences not possible from qualitative research. However, the use of AI as a research method to inform the project (re) design raised ethical questions such as those relating to data privacy and informed consent. The framework produced in partnership with Loughborough University addresses these and provides researchers with guidance on how to design, analyse and report on research findings.

Discussion/Implications For The Field:

Increasingly, SBCC practitioners will be deploying social media to reach target audiences and exploring new research techniques such as AI to understand audiences effectively. Researchers need to address the ethical challenges underpinning digital research in general, and that using AI in particular. BBC Media Action will share the framework it put together to navigate the ethics of AI and machine learning in research.

Abstract submitted by: 

Shams  Tazi - BBC Media Action

Elise Ketelaars - BBC Media Action