In September 2024, Verian and Faculty co-hosted an event to discuss the role of Artificial Intelligence (AI) in accelerating research.
Event details
Hosted in Westminster, London, this event brought together senior leaders from across government and public bodies to explore the practical implementations of data and AI-enabled approaches to both qualitative and quantitative research.
Speakers and panellists included:
- Sir Ian Diamond, National Statistician of the UK Government Statistical Service
- Alison Kilburn, Director of Analysis at Department for Science, Innovation and Technology
- Ed Bearryman, Deputy Director Digital Transformation at Government Communication Service
- Ben Humberstone, Head of Population Studies at Verian UK
The event was introduced by Craig Watkins, UK CEO at Verian and the panel session moderated by Nijma Khan, Director of Government and Public Services at Faculty.
The panel discussed policy problems where AI could help to shed light, practical applications, potential concerns and barriers, and opportunities for using AI and novel data sources in research.
The potential impact of AI in social research
More so than ever, governments and policy makers need access to high quality, accurate, timely and representative research and data to inform policy development, understand public service user’s needs and deliver operationally. Whilst research has traditionally involved a mixture of qualitative and quantitative research techniques such as large-scale panel surveys and indices, public consultations, and focus groups and interviews, new technologies and novel sources of data present exciting opportunities for advancing research in public policy.
Potential benefits of AI in social research are:
- Using AI and novel data sources for social research is part of the ‘future of evidence’, leading to new types of insight and analysis answering previously unexplored research questions, more rapidly, at a larger scale and with greater richness.
- Accessing novel data sources such as app-based data, social media and other data created by new technologies provides opportunities for new and more granular analysis for decision makers.
- Research with AI and Machine Learning will lead to more in-depth and useful understanding of sub-groups of the population, incorporating novel data sources can help reach seldom heard voices or harder to reach, more marginalised communities.
- AI has made the linkage of disparate datasets at speed and scale possible; by connecting data sets, researchers and policy makers can look at the holistic impact of particular issues on people and build a richer picture of people’s experiences.
- New analysis is possible through use of Natural Language Processing and other applications of Machine Learning, making sense of large unstructured datasets.
- More transparency is possible with differential privacy techniques to share and analyse new data sources.
- Using Large Language Models to process survey and interview responses drastically increases the speed and scale of research – if used safely and responsibly. The Smart Matrix Analyser, developed by Verian and Faculty, is an example of the exciting possibilities of AI in this area, automating the coding and analysis of qualitative data.
Considerations
The two discussed at the event were:
- The critical role of humans in conducting research
- Ethical considerations when using AI
Humans play a critical role in research, which cannot be replaced by AI
The continued critical role of human expertise and experience to understand the research question, the study population, and the models used is important. Without this, we risk simply pointing powerful computers at random problems, leading to misinterpretation and misunderstanding of the data.
AI cannot overcome the fact that when investigating a social issue, one needs to know and understand that social issue. People are necessary for data interpretation as they can understand the context and nuance. We should see AI as enhancing and improving the work of researchers, not replacing it.
Ethical considerations of AI in research
Considerations discussed by the panel include:
- Skills – Training people on how to use AI tools is vitally important. In particular, ensuring people have the skills to use and interpret data in a responsible way.
- Transparency - There are issues surrounding the security and ethics of data sources especially on sensitive topics such as health data. To use AI and novel data sources responsibly, honesty and transparency is important.
- Bias – we must be conscious of biases in the algorithm as well as in the data sets. Human understanding of the context surrounding the data is important to acknowledging and reducing bias.
In conclusion
Embracing AI and novel data sources can support timely and inclusive research, in turn informing better policy decision-making. Whilst the responsible use of AI is leading to new types of analysis informing previously unanswered questions, at greater speed and scale, human understanding and expertise will continue to be vital to address the research question. To use AI responsibly, researchers and policy makers need to be well informed, thoughtful and transparent about its use.
Additional resources
- Context model: Read the case study
- DCMS engagement: Read the case study
- Fellowship project: Read the case study
- Smart matrix analyser: Read the case study
To find out more, get in touch: