Generative AI policy – what staff need to know
16 hours ago
Carmarthenshire County Council has introduced a new Generative Artificial Intelligence (GenAI) Policy to help everyone use AI tools safely, ethically and in line with our legal and organisational responsibilities.
The policy applies to all employees, agency workers, volunteers, elected members and external contractors using council digital systems.
Why we have a GenAI policy:
Many of us are now using AI tools such as ChatGPT, Microsoft Copilot, Claude, Gemini, Perplexity AI and others to support everyday tasks like:
- Writing or refining emails
- Summarising long documents
- Creating ideas or draft content
- Researching topics
These tools can be extremely helpful, but they also come with important risks, including:
- Inaccurate information (“hallucinations”)
- Bias or misleading outputs
- Data protection issues
- Security risks if confidential information is entered
The new policy has been created to ensure that AI tools are used safely, responsibly and to the benefit of the organisation and staff.
It ensures that our use of AI is always:
- Ethical
- Secure
- Compliant with GDPR and the Data Protection Act 2018
- Aligned with our information governance standards
Which AI tools should I use?
Where possible, staff should use Microsoft Copilot Chat in Teams or Microsoft Edge rather than public generative AI tools such as:
· ChatGPT (OpenAI)
· Claude (Anthropic)
· Gemini (Google)
· Perplexity AI
· Grok AI
· Deep AI
Copilot Chat operates within the Council’s secure Microsoft 365 environment, meaning prompts and outputs are handled in line with our information governance, security and data protection controls.
This makes it a safer option for day-to-day work such as drafting content, summarising information or refining ideas.
That said, the Generative AI Policy is not intended to restrict innovation or prevent the use of wider free online AI tools. Staff may continue to use public tools where appropriate, provided they do not input personal, sensitive or confidential information, apply human judgement, and follow the principles set out in the policy.
AI learning resources for all staff:
To help staff build confidence in using emerging technologies, our new AI Essentials badge, a series of three AI eLearning modules, is available on Thinqi now. These modules are designed to provide clear, practical guidance on artificial intelligence, whilst helping you to explore how AI works, how it can support day-to-day work across the organisation, and how to use AI tools responsibly, safely, and effectively.
Building on this, an additional module, Introduction to Copilot, will be launching soon. This upcoming course along with further learning opportunities, further strengthening staff skills and confidence helping everyone make the most out of AI tools.
Ø