How we developed ACV-CSC’s AI policy - and what it is helping us do using AI
ACV-CSC is leading the way in AI adoption, and one of the few unions to create their own AI policy. Here's the story of how they developed their policy.
Johan Van Baelen
April 29, 2025
At ACV-CSC, Belgium's largest trade union/federation with 1.5 million members, we knew we needed to get AI right. Not just for our members, but for our staff too.
When we started seeing colleagues experiment with AI tools like ChatGPT, we realised we needed clear guidelines to help everyone use AI safely and effectively.
Starting with what matters
Rather than rushing in, we took time to consult widely across our union. Through surveys and in-depth discussions with our decision-makers, we explored what really mattered when it came to using AI in our work. These conversations helped us identify four core principles that now guide everything we do with AI.
First and foremost, we believe humans must stay in control. This means any output from AI needs to be verified by our staff, who remain responsible for the final results. We treat AI-generated information like we would a first draft - useful, but requiring human oversight and verification before it can be trusted.
Safety and privacy are also absolutely crucial. We've learned to be selective about which AI tools we use, choosing only those that keep our data secure. This careful approach is why we now recommend Microsoft Copilot over ChatGPT for our staff - it offers better protection for sensitive information.
We're also committed to being open about how we use AI. Whether we're communicating with members or developing internal systems, we believe in being transparent. This led to some fascinating discussions about finding the right balance - how to be clear about AI use without overwhelming people with technical details.
Finally, we're mindful about when we deploy AI. Given the significant energy and resources it requires, we only use it where it brings clear benefits. If traditional tools can do the job just as well, that's what we stick with.
You can see the latest version of our policy below:
|
|
AI technology moves incredibly quickly, and we've learned that our policy needs to keep pace. We regularly want to review and update our guidelines based on what we're learning from how our staff use AI in practice. It's a constant process of refinement and improvement.
 Do you work in a union and are you looking to learn more about AI? With our self-paced training you can start your learning journey today. | Unions21’s Responsible AI Fundamentals for Union Staff courses go beyond the hype around AI, to help you understand the real opportunities and challenges of AI for you, your team and your union. Register today for our instant start self-paced course and join the only training course focused on helping union staff understand AI for their work. |
From policy to practice
Having these clear guidelines has given our teams the confidence to experiment with AI in ways that really make a difference. We're now testing Microsoft Copilot across the union, helping staff work more efficiently. One of our projects has been developing an AI system that helps process 600 member emails daily, identifying urgent issues and key questions that need attention. Teams are also exploring practical applications like translation and video subtitling.
In this video, Steven Willocx (also of ACV-CVS) describes the AI experiment they're running, with a system that helps staff process the hundreds of emails that arrive in a shared email inbox every day.
Advice for other unions
If you're thinking about developing your own AI policy, I'd encourage you to start by talking to your people about what matters to them. Keep your principles clear and practical, and make sure your guidelines help rather than hinder. Be prepared to update as technology evolves, and look for early opportunities to show how AI can add value to your work.
Most importantly, remember that many of your staff are probably already using AI in some way. A good policy helps them do this safely and effectively, rather than trying to control or restrict their use completely.
By setting clear guidelines first, we've been able to help our staff take advantage of AI's benefits while minimising some of its risks. This has been important in building trust and ensuring AI is really there to help us be stronger.
Video: I explain my advice for unions starting their AI journeys in the recent webinar: AI in action - How unions are using AI today
Johan van Baelen is a staff member of the recruitment and communications department at ACV-CSC, Belgium's largest trade union. He works alongside Stephen Willocx, head of AI in the business and IT department, to develop responsible approaches to AI use across the union.
Transparency: this blog was written with the support of AI, based on my contributions to the webinar.
Reply