This website uses cookies

Read our Privacy policy and Terms of use for more information.

A week ago we launched the first ever global survey of trade union staff on AI. We want to know their experience, hopes and fears with AI.

The survey will be open for a few more weeks, but I wanted to get a sense of what might be in it. And show how AI can help do this task quickly, including some categorising data in a way that is really useful.

In the survey, I specifically didn't bother to worry about creating a drop down list for "Union", "Job role" and "Country" because I knew I could get AI to help me – and therefore make it easier for users who don't like to spend time searching long lists of options.

Watch the video to see how I quickly analysed the results:

How AI can enhance union surveys: Key opportunities

Unions love surveys. There’s a lot of members, and lots of views to take account of. However they often face common challenges - from survey design to analysis of results. AI can help at every stage – that it is really important to understand its limitations and how to use it responsibly.

Survey design and testing

Data cleaning and categorisation

Analysis and insights

Time-saving automations

Generate question variations to test different phrasings

Standardise free text responses into consistent categories

Quickly summarise key themes from open-ended responses

Create template emails for survey reminders

Help identify potential bias in question wording

Clean and format messy data

Identify patterns and correlations across different questions

Generate initial drafts of survey reports

Suggest additional questions based on survey objectives

Flag potential data quality issues or inconsistencies

Generate initial hypotheses for deeper investigation

Prepare data visualisations of key findings

Make the language simpler and more plain, finding different ways to word complex questions

Convert varied location formats into standard geographical categories

Create draft summaries of findings for different audiences

Format results for different outputs (presentations, reports, social media)

Important safeguards when using AI with survey data

AI can potentially do a lot, as the table above shows. However it is important to be honest: sometimes using tools like ChatGPT actually create more work than doing it yourself, particularly in terms of data analysis. They can get things wrong (in particular anything involving maths), make things up, or get confused in a number of different ways. This is particularly the case with the free versions of the tools

And, at other times using AI is not responsible. To be responsible in your use of AI tools you should:

  • Never input personal or identifiable information – the survey I added was anonymous and did not contain names, emails or another other personal data.

  • Always verify AI-generated analysis against the raw data. In this case, I just wanted to get a sense of the overall numbers, so the response Claude.ai gave me was fine. But if I were doing serious research I’d check everything it says for myself.

  • Keep human oversight of question design and final conclusions.

  • Be transparent about AI use in your survey processes.

Complete the survey on AI in trade unions

AI is changing work. As unions, we've faced many changes before - and helped workers navigate them.

Now we need to look at how AI is changing our own work, as staff working in unions. In 2024, Unions 21 spoke to hundreds of union staff about AI. We learned many are already using AI tools to help with their work.

Some staff talk openly about using AI. Others keep it private - not telling colleagues or leaders they use it.

To help workers deal with AI at work, we first need to understand how we use AI ourselves.

This study is the first to look at AI use across unions worldwide. Your answers will help unions develop practical and ethical ways to use AI responsibly.

Add your voice to this important research. Already 450+ union staff from 25 countries are sharing their experiences.

Reply

Avatar

or to participate

Avatar

Nick Scott2 years ago

I promised to include the prompt I used in the video, so here it is (1 of 2)

Act as a data analyst. I've attached the interim results of a survey of union staff on their thoughts and experiences with AI. I'd like you to bring deep knowledge of generative AI and trade unions to analyse these results. First, I want to get a sense of who has completed the survey. To do this we might need to do some tidying and collating of data in different fields, as currently the field for job role, union and country is free text. Use your knowledge to understand and collate these together into individual unions, departments and countries. For example, Northern Ireland, UK and United Kingdom are all one.

Avatar

Nick Scott2 years ago

2/2 Next, I want to get a sense of the overall top-line narrative from the results so far. What might we take away from the data so far about the level of experience union staff have with AI, what their key hopes and fears are and which ways they are using AI. It would then be interesting to look at whether there are any particular segments - unions or union departments - where there are different results to the norm. Finally, it would be useful to pick out some of the most interesting insights and comments from the free text fields, to add colour to the results.