A year ago, we launched the first global survey of AI use in unions. Since then, through Unions 21 I have run training for hundreds of union staff, from London to Los Angeles, from Brussels to Sydney.
I've heard so many stories of union staff using AI. How they are using it. Why they are using it. Where they find hope. Where they fear its impact.
Some of this AI usage is open… but much of it is hidden. Hidden from other staff, and from union leaders.
Secret usage of AI isn't just a problem for trade unions — it is happening in lots of organisations.
But unions have a special responsibility. If we are to stand up for workers in the AI age, we must first understand our own relationship with this technology.
What we learned from last year's survey
Last year's survey was the first of its kind. We heard from 314 union staff, across 116 unions, in more than 50 countries. Here is what they told us.
Union staff are already using AI… mostly without training
The headline finding is stark: 80% of respondents have used AI for at least one work task. The most common uses are developing text (67%), transcribing or summarising content (61%), and research or idea generation (42%).
But here's the problem: two-thirds of respondents (67%) have never had any AI training. And of those untrained staff, 70% are using AI anyway.
This is exactly the pattern we see across sectors – staff are experimenting with powerful tools without guidance on how to use them safely or effectively. For unions handling sensitive member data and confidential casework, this gap between usage and training should concern us all.
Staff are cautiously hopeful – but real fears persist
When we asked union staff how they feel about using AI in their daily work, the picture is one of cautious optimism. 45% feel positive, 31% are neutral, and 23% feel negative – an average sentiment score of 5.2 out of 10.
Their top three hopes tell a clear story:
Productivity: getting more work done efficiently (cited by 67% of respondents)
Union impact: stronger advocacy and better member engagement (45%)
Professional development: learning new skills and staying relevant (43%)
But so do their top three fears:
AI doing harm: perpetuating biases or creating disparities (51%)
Privacy and data security: putting member data and confidentiality at risk (47%)
Loss of human touch: AI diminishing personal interaction and empathy (47%)
These fears are not abstract. They reflect the real tensions that union staff navigate every day. To use AI you have to balance the hope of efficiency with the reality of human relationships at the heart of union work.
Unions as organisations are not keeping pace
Perhaps the most important finding is the gap between individual staff adoption and organisational readiness.
While 80% of staff are using AI personally, only 8% say their union is "active" on AI for internal operations. Another 40% describe their union as "exploring," but 36% say their union is "not engaged" at all.
The policy picture is equally concerning: 58% of respondents either have no AI policy or aren't sure if one exists. Only 12% say their union already has a policy in place.
Staff know what they need. When asked what would help, the answers were clear: training workshops and courses (75%), case studies from other unions (61%), peer connections to discuss AI with colleagues in similar roles (49%), and access to AI experts (47%).
Why we're doing this again
These findings confirmed what many of us suspected — but they also revealed just how wide the gap has grown between where union staff are and where their organisations need to be.
That's why I am repeating last year's survey. A second year of data will help us understand whether these patterns are shifting — and what unions need to do next.
I ask you to do two things today:
Share the survey with your colleagues so that we can get an accurate picture.
The survey is available here: https://unions21.fillout.com/ai-survey-2026
Thank you for your help!
Nick Scott
Director, the Centre for Responsible Union AI
