Last week we attended both the Social Research Annual Conference and a half-day conference on how AI is changing social research, jointly hosted by GSR, SRA and MRS. After two days in sweltering central London, we've finally cooled down enough to reflect on some of our key takeaways—or perhaps heat-induced musings.

Social researchers, as everyone, are obsessed with all things AI: keen to learn what others are doing and implementing in this space; keen not to miss out on the potential benefits; and even keener to think through the ethical implications AI technology has for participants.

Talk of AI-generated prompts, synthetic data and the ability to code vast numbers of open-ended responses left us wondering whether AI might encourage researchers to ask more questions and generate more data simply because we can. Synthetic respondents won't suffer survey fatigue, coding costs won’t be such a barrier. However, responsible use of AI means that we still need to consider asking only the questions that genuinely add value – and not just adding questions in, simply because we can. We were surprised there wasn’t more discussion of the environmental impact of AI. Perhaps it was too uncomfortable to consider the climatic costs of these technologies while sitting through yet another recording-breaking heatwave.

These concerns aside, we came away thinking it's okay to use AI in parts of the research process—provided we don't build it into every stage, we're transparent with clients and participants about how it's being used, and we're confident about data security and how participant data will be stored and used.

We were also reassured that nobody seems to be advocating for AI to replace in-depth interviewing. Data presented by Glaut[1] showed that AI-moderated interviews scored lower than human moderation when it came to building emotional connections with participants. Encouragingly, even they recommend human interviewers when emotional connection or deep relational sensitivity is central to the research.

Our final takeaway was that social researchers are great storytellers. They know how to bring data to life, connect with audiences and sprinkle presentations with just enough humour to remind us that people still sit at the heart of good research—even if part of that job now involves sense-checking AI prompts and summaries.

[1] https://www.glaut.com/glaut-research/a-comparative-study-of-ai-moderation-vs-human-moderation-using-biometrics-by-curtin-university

Get in touch

A hot take on AI from the summer conference season