Community survey

What we learned from the first AI-assisted coding meetup survey

At our first meetup my intention had been to go around the room and ask people to briefly introduce themselves and talk about their experiences with AI-assisted coding. However, the day before, when it became apparent that we might have 50+ attendees, I decided to hastily put together a short survey that we could run during the meetup. You can explore the results of that survey below.

Disclaimer: this is not a formal piece of research, so results should be taken with a grain of salt! One respondent was removed from the dataset due to them being potentially identifiable via Dept/College association.

View responses by All respondents Role Dept/College Required to write code

What is your experience with AI-assisted coding?

Single-choice responses as a share of each group.

Which of these tools have you used?

Multi-select tool families; percentages can add to more than 100%.

Which of these statements reflect your opinions of AI-assisted coding?

Multi-select opinion statements, shown as percentages of each group.

What would you most like to gain from these events?

Multi-select up to 3, shown as percentages of each group.

Adoption and tools

Almost 70% of respondents who are required to write code as part of their role said that they already use AI-assisted coding tools on a regular basis (multiple times per week). One of the aims of these meetups is to support people who are unfamiliar with these tools to get started, so it was encouraging to see that we had a number of attendees who had never used AI-assisted coding tools before. Copilot and Claude Code shared the top spot for most commonly used tools. Whilst Copilot can be accessed for free by staff and students via ELM or GitHub Education, it is notable that Claude Code is a paid product that is not currently available via University channels.

Cautious optimism

In my conversations with colleagues I encounter a lot of scepticism around the use (and impact) of generative AI tools. So I was somewhat surprised by the extent to which attendees' opinions towards AI-assisted coding leaned positive. 70% of respondents stated that "It improves my productivity" and 64% said "It makes my job easier". Meanwhile, concerns for environmental impacts were shared by 40% of respondents, and 48% said they "fear being left behind if I don't adopt it". Concerns over AI tools impacting job security were only expressed by respondents who are required to write code as part of their roles (27%). 22% of respondents said that they don't trust AI-assisted coding tools. This response rate didn't vary much by the respondent's Dept/College or whether they are required to write code as part of their role, but mistrust was higher among academics (36%) than among professional services staff (22%).

What this means for future meetups

There is a clear desire for these meetups to provide opportunities for people to learn new skills related to AI-assisted coding, and to share knowledge and experiences with others. The pace of development in this space is difficult to keep up with, so there seems to be a strong case for monthly meetups to provide an opportunity to disseminate information about new tools and approaches. The free-text suggestions for future meetup topics were practical and varied, but several themes came through clearly:

Getting started and best practices for the use of AI coding tools.

Comparing tools, models, and IDE setups.

Agentic coding, MCP servers, and task automation.

Security, data sovereignty, policy, and transparency.

Testing, verification, documentation, reuse, and maintenance.

Tracking usage/cost and environmental impacts.