programme

As part of the [AI]volution theme track we host a hands-on workshop on Tuesday the 14th in the evening (and, in case of sufficient interest, once again on the 15th) on how Large Language Models (LLMs) can be responsibly used to analyse qualitative data.

With the impressive capabilities of modern LLMs such as Claude, ChatGPT or Gemini, AI has become a tempting alternative for data-related tasks that were highly time-consuming to do manually, or required crowd-sourcing via platforms such as Prolific. However, there are a few important qualifications to be made. Research data may contain sensitive information, so sending it to external servers (e.g., OpenAI, Meta, Google) is often undesirable or, in some cases, legally restricted. Large commercial LLMs are notorious for their energy and drinking water consumption and often employ business models that run against academic interests.

​

In this workshop, we therefore provide a practical alternative: using small, locally runnable language models to analyse, classify, and process qualitative data, such as responses from open-text fields in a survey, while keeping all data on your own machine and using a fraction of the energy.

​

The workshop is open to all EHBEA participants. If you’d like to code along, please install Python (https://www.python.org/downloads/) and Ollama (https://ollama.com/download/) beforehand. Support will be available on the day for anyone who needs assistance with setup. You are also very welcome to join the workshop without coding along.

​

We look forward to seeing you there!

​

Lennard Froma & the Local Organising Team