For decades, computer science research was purely technical – the algorithmic architecture, model training, and data pipelines were only mathematical and engineering challenges. Because of this, they were kept only to university labs and corporate R&D facilities. However, as artificial intelligence (AI) and data-driven systems have made their way into healthcare, criminal justice and daily human communication, these technical systems no longer hold up. The problem? The real-world harms of unchecked AI have exposed a critical flaw in traditional computer science – that systems built for people without including them can inflict unintended societal harm.
The solution? The researchers who are redefining the boundaries of technology research by making creative, participatory, and interdisciplinary methodologies that embed citizen engagement and co-creation into the building phase of computer science research.
Citizen Engagement
Leading this change is Dr Keeley Crockett, Professor of Computational Intelligence at Manchester Metropolitan University. Crockett’s work shows how computer science can break free from traditional lab settings by using community co-production as a primary scientific method.
Historically, computer scientists treated human users merely as subjects who generated data points or validated interfaces. However, Crockett’s paradigm flips this dynamic through projects such as PEAs in Pods (Public Engagement for Data and AI Research) and the People’s Panel for AI. Rather than bringing a fully formed machine learning model to a community for testing, these initiatives involve citizens (particularly those from traditionally marginalised backgrounds) in the co-creation of the research questions, ethical frameworks, and system boundaries themselves.
Through public engagement workshops that help break down complex computational concepts like ‘fuzzy logic’ and ‘neural networks’, agency goes back into the hands of the citizens. Through different exercises, participants articulate how their personal data should be handled, evaluate the acceptable risk thresholds for algorithmic decisions, and challenge assumptions that computer scientists might take for granted. This direct involvement transforms qualitative public experience into structural parameters for algorithmic design, ensuring that ethical AI research directly reflects public values and cultural nuances.
Key Creative Research Methodologies in Computational Science
These participatory frameworks have made way for more creative methodologies in computational and social data science. These approaches bridge the gap between abstract code and actual human experiences.
- Speculative Design. Speculative design uses mock scenarios and storytelling to explore the potential future impacts of emerging technology. Instead of asking “What do we want from an AI tool today?”, researchers present participants with fictional future news stories, speculative Terms of Service agreements, or mock interactive devices. The reaction that occurs as part of these scenarios shows underlying fears, structural vulnerabilities, and ethical boundaries that could otherwise remain hidden.
- Participatory Data Auditing. Rather than relying only on automated benchmarking datasets to test for bias, citizen data auditing takes the power away from the tech companies and shifts it to non-experts who interrogate the datasets directly. The people involved review the category labels, proxy variables, and historical training data used by predictive systems. By doing this, hidden biases are highlighted in spaces that developers might overlook. Overall, actual lived community knowledge is made into a diagnostic tool for algorithmic audits.
- Living Labs and Multi-Modal Dialogic Frameworks. Living Labs establish open, real-world research ecosystems within cities and neighbourhood centres. By combining physical installations, natural language interfaces, and conversational agents, researchers create interactive environments where community members engage with data technology in their day-to-day settings. This methodology replaces artificial laboratory conditions with natural, context-rich interactions, generating higher-fidelity data on how humans actually trust, reject, or negotiate with automated systems.
Is There a Downside?
While creative participatory research does have many benefits for making AI development more ethical, integrating qualitative practices into a quantitative field like computer science introduces some unique challenges:
- The Epistemic Gap. Standard computer science prioritises metrics like accuracy, precision, recall, and efficiency. Participatory methods prioritise trust, systemic fairness, context, and human lived experience. Adding these qualitative insights into quantitative optimisation models requires new computational frameworks, such as fuzzy systems that can mathematically encode human ambiguity and linguistic uncertainty.
- Power Asymmetries. Traditional technology research often positions the academic or engineer as the expert and the community member as a passive participant. Creative methods require active power-sharing, where citizen contributors are recognised and compensated as genuine co-researchers whose expertise shapes the core project trajectory. Whilst not typically a negative, it might present a challenge to academics used to their old ways of working.
- Institutional Constraints. Academic funding cycles, peer-review norms, and publication formats are traditionally structured around quick lab experiments rather than the slow, trust-building timelines necessary for good community engagement. Moving these methodologies into the mainstream requires structural shifts in how research impact is funded and evaluated – which takes time!
The Future of Research
The work of pioneers like Dr Keeley Crockett highlights a necessary evolution in modern computer science. As artificial intelligence becomes deeply interwoven with daily life, the traditional boundary between the technical system and its social environment has collapsed.
Creative participatory methods show us that citizen engagement is not just an ethical add-on or a public relations exercise, but instead, it is a rigorous, essential research methodology. By incorporating public co-creation, speculative design, and community-driven data audits into the early stages of computing research, computer scientists can build systems that are technically sound, democratically accountable, and grounded in human values. What are your thoughts on this change?
