Using an LLM to Find Themes in an Academic Career

Brief recap

In an earlier post, “Same prompt, four AIs: why answers aren’t always the same”, I looked at what happened when the same prompt was given to four different LLMs. Unsurprisingly, perhaps, the answers were not identical. That raised an interesting follow-on question: what was the prompt actually trying to do?

The answer is that it was trying to support a form of thematic analysis. In that case, the object of analysis was an academic profile, including my Google Scholar profile. The version I discuss here is a modified prompt, used with Claude.ai, where I uploaded a CV and asked the system to find other appropriate public resources connected with a named person and institution. The aim was not simply to summarise the CV, but to identify visible and less visible themes across a whole career.

Why use thematic analysis?

Thematic analysis is widely used by academics, especially in qualitative research. Braun and Clarke’s well-known paper, Using thematic analysis in psychology (V. Braun and V. Clarke, Qualitative Research in Psychology, 3(2), 2006, pp. 77–101, DOI: 10.1191/1478088706qp063oa), sets out a staged approach involving familiarisation with the data, coding, searching for themes, reviewing themes, defining and naming themes, and producing the final account.

For a personal academic reflection exercise, this is useful because a career is rarely just a list of outputs. A standard CV summary might say what roles someone held, what they published, or which grants and activities they were involved in. A thematic analysis asks a different sort of question: what patterns seem to run through this work?

Some of those patterns may be semantic themes: themes visible on the surface of the material. For example, a CV might show repeated work around computing education, employability, widening participation, or public engagement. Others may be latent themes: deeper or more interpretive patterns that are not stated directly but may be inferred across the material. For example, the profile might suggest a continuing concern with access, community benefit, interdisciplinarity, or building bridges between formal education and informal learning.

What the prompt does

The prompt asks the LLM to analyse an uploaded CV alongside other available and appropriate resources on a named person and their institution or organisation. It sets the scope as the whole career, asks for all available data to be considered, and requests both semantic and latent themes.

One useful feature is that the prompt asks the LLM to ask questions before beginning the analysis. This is not a minor detail. The questions are part of the value, because they force clarification around purpose, scope, sources, and constraints before the tool starts interpreting the material.

In my case, the output was useful for reflection. For someone with an interdisciplinary publication record and non-traditional academic activity, such as setting up Code Clubs, leading STEM-related groups, or undertaking public and community roles, the thematic lens can surface connections that may not be obvious from a chronological CV.

Using the output cautiously

This is not about outsourcing judgement to an LLM. It is better understood as a practical but imperfect tool for structured reflection. The themes it produces should be treated as provisional suggestions, not objective truths.

There are several reasons for caution. First, the analysis is only as good as the sources. If the CV omits important teaching, leadership, collegiality, public engagement, or service work, the LLM may not identify those themes. That absence can still be useful: it may show where the CV needs to capture more of the career story.

Second, if the LLM searches for public sources, it may find irrelevant, incomplete, or incorrect material. Third, latent themes are interpretive. They can be helpful, but they can also be flattering, speculative, or biased. The user needs to check whether each theme is actually supported by the evidence.

There are also privacy and ethical issues. Uploading a CV to an AI tool should be done with care, especially if it includes personal data, unpublished work, confidential projects, student information, or institutional material.

What could be improved next?

A useful future improvement would be a two-stage process. First, ask the LLM to list the sources it has found and ask the user to confirm which should be included. Only after that should it carry out the thematic analysis. This would reduce the risk of spurious sources being used.

Another improvement would be to ask the LLM to provide evidence for each theme, ideally by quoting or citing the relevant CV sections, publication titles, abstracts, institutional pages, or other sources. Future versions of the prompt could also specify extra sources directly, such as an ORCID profile, Google Scholar page, institutional profile, or selected public engagement evidence.

How I would use the output

I would use the generated themes as a reflective aid; comparing them with my own understanding of my career, looking for gaps, challenging anything that feels too neat, and considering whether the CV itself needs rewriting to make important themes more visible.

Different LLMs may produce different results from the same prompt, as the earlier post showed. That is not necessarily a weakness, but it is a reminder that the output is an interpretation, not a verdict.

The prompt

Here is the lightly corrected version of the prompt I used:

From the uploaded file and the other available appropriate resources on <person’s name and institution/organisation>, carry out a thematic analysis. Before doing the analysis, please ask any questions you need, then do the analysis.

Please use the steps in Braun and Clarke (V. Braun, V. Clarke, Using thematic analysis in psychology, Qualitative Research in Psychology, 3(2), 2006, pp. 77–101, DOI: 10.1191/1478088706qp063oa) to do this.

Constraints

Scope: Whole career
Depth of data: All available data, including publication titles, abstracts, and other appropriate sources
Analytical lens: No specific research question
Target themes: Look for semantic themes and then provide a separate analysis of latent themes

Please ask any clarification questions before starting the analysis.

Try adapting it cautiously, and remember: the value is not just in the themes the LLM suggests, but in how those themes help you reflect, revise, and question your own career narrative.

All opinions in this blog are the Author's and should not in any way be seen as reflecting the views of any organisation the Author has any association with. Twitter @scottturneruon