Completing my Independent Research Project (IRP) was one of the most rewarding experiences of my degree. I’ve built my career in the blockchain industry, with a growing interest in artificial intelligence. This project allowed me to combine my experience and interest while developing valuable research and technical skills along the way.

My experience

When choosing a topic, I wanted to work on something that felt relevant both academically and professionally. Working for a foundation that grows and develops a Decentralised Autonomous Organisation (DAO) provided me with a clear understanding of the challenges they face.

At the same time, artificial intelligence is rapidly transforming the way people access and process information, leading me to an important question: could AI help improve participation in decentralised governance?

What began as a simple idea quickly evolved into a much larger research challenge, and I quickly learnt the importance of narrowing its scope. As with all research projects, there were countless directions that could be explored. Throughout my Research Proposal module, and with the help of my supervisor, I focused the project on three specific areas: proposal summarisation, participation analytics, and the ethical implications of AI in governance.

The project pushed me far beyond my existing skill set, spending significant time reviewing academic literature, learning new analytical techniques, exploring which algorithms were most appropriate, and building a computational pipeline in Python. While I had previous experience working with data, conducting independent academic research required a different mindset. Justifying decisions and supporting any claims with evidence was key, while ensuring any limitations were clearly acknowledged.

Justifying decisions and supporting any claims with evidence was key, while ensuring any limitations were clearly acknowledged.

One of the most challenging aspects was balancing technical implementation with academic rigour. It was not enough to build a working solution; I also needed to explain why the methodology was appropriate and how the findings related to existing research.

Perhaps the most valuable lesson was learning to embrace uncertainty. Many of my initial assumptions were challenged by the data. Rather than reinforcing my expectations, the research often revealed more interesting and nuanced insights. This taught me that successful research is about following the evidence wherever it leads.

What I achieved

My project examined how AI could support governance within a real-world DAO environment. Using RootstockCollective as a case study, I analysed over 100 governance proposals, nearly 4,000 votes, and participation data from more than 450 unique wallets.

To explore whether AI could make governance more accessible, I developed a Natural Language Processing (NLP) summarisation pipeline to generate concise proposal summaries, aiming to reduce the cognitive effort required to understand complex proposals while preserving their original meaning.

The analysis showed that AI-generated summaries successfully reduced proposal length and informational complexity. However, findings also revealed that proposal length alone was not a significant barrier to participation. Longer proposals often attracted similar or even higher levels of engagement, suggesting that proposal relevance and perceived importance may play a greater role than information volume alone.

One of the most significant findings was the degree of concentration of voting power. Using inequality metrics such as the Gini coefficient, I found that a relatively small number of participants controlled a substantial proportion of total voting power, raising important questions around the design of token-weighted governance systems.

Beyond summarisation, I applied AI-driven analytics to investigate voting behaviour. This analysis uncovered several interesting patterns. Participation was highly selective and often concentrated around proposal types, particularly grants and builder funding. Voting activity was also strongly event-driven, with engagement clustering around specific governance cycles.

One of the most significant findings was the degree of concentration of voting power. Using inequality metrics such as the Gini coefficient, I found that a relatively small number of participants controlled a substantial proportion of total voting power, raising important questions around the design of token-weighted governance systems. However, the analysis was unable to separate large individual token holders from delegates, who often represent multiple voters and drive decentralisation.

The project also explored the ethical implications of introducing AI into governance processes. While AI has the potential to improve accessibility and transparency, it may also raise concerns about algorithmic bias and reliance on centralised AI infrastructure. These findings highlighted the importance of ensuring that AI remains a support tool rather than replacing human decision-making within decentralised communities.

Looking back

This project allowed me to combine academic research, artificial intelligence, blockchain technology, and governance into a single piece of work. More importantly, it helped me develop skills that extend far beyond the project itself, including critical thinking, data analysis, technical implementation, research design, and evidence-based decision making.

For future students beginning their IRP journey, my advice is simple: choose a topic you genuinely care about and be prepared for your research to challenge your assumptions.

For future students beginning their IRP journey, my advice is simple: choose a topic you genuinely care about and be prepared for your research to challenge your assumptions. The final report is important, but the knowledge and skills you gain throughout the process are what make the experience truly valuable.

• LinkedIn