Aim of the policy
The aim of the policy is to ensure that the use of artificial intelligence (AI) within UMFÍ is responsible and ethical and supports the organisation’s role. AI should be a tool to lighten the workload, increase efficiency and encourage innovation, but never at the expense of human interaction, security or data protection.
UMFÍ’s principles
The human element comes first: AI never replaces personal relationships, community and the work carried out within UMFÍ. The technology is a support tool, not a substitute.
Transparency: We are open about when and how we use AI, especially if it is used in creating content or in decision-making.
Data protection and security: We take the utmost care with security and comply with data protection law (GDPR). Data about children, young people and volunteers is never entered into open AI systems.
Responsibility: Staff and volunteers are always responsible for the content and results that AI helps create. AI must never be trusted blindly. Users of AI tools shall always use them responsibly and think critically.
Permitted and recommended use of AI
UMFÍ staff are encouraged to use AI responsibly, for example to help with the following:
- Brainstorming and planning: Getting ideas for events, projects, programmes or headlines.
- Administration and office work: Proofreading text, translating content, summarising meeting minutes (without sensitive personal data) or drafting emails.
- Content creation: Help with creating promotional material for the website (umfi.is) and social media, as long as the content is reviewed and fact-checked by people.
- Data processing: AI can help with processing data, which can both save time and help with analysis.
The AI tool Claude has been approved, and all staff have a work account for Claude linked to their work email address.
Restrictions and prohibitions (the red lines)
Registration and work email addresses: When AI tools are used on behalf of UMFÍ, or to create content for the organisation, users shall always use their account based on the organisation’s email address (an @umfi.is address).
It is not permitted to use personal email addresses or private accounts to work with the movement’s data or content.
Sharing personal data: It is not permitted to enter names, national ID numbers, email addresses, photographs or other personally identifiable information about participants, volunteers or staff into open AI tools (e.g. ChatGPT, Copilot, Midjourney). If an exception is required, an application must be made, and appropriate measures will be taken to ensure that the data does not leave UMFÍ’s systems.
Factual errors and “AI hallucinations”: AI can produce incorrect information. It is not permitted to publish content on UMFÍ’s channels unless a person has read it and confirmed that the information is correct. The user of the AI is always responsible for the content they publish, and therefore for the facts being correct. AI shall always be used with caution and critical thinking.
Illustrations and authenticity: When AI is used to create images, it shall be ensured that they show diversity and reflect UMFÍ’s values. If an image is clearly created with the help of AI and could mislead people, it shall be specially labelled (e.g. “Image: AI/UMFÍ or the user’s name”).
Implementation, training and review
Training: UMFÍ places emphasis on training its staff and board members in the safe and practical use of AI.
Review: As the technology advances rapidly, this policy shall be reviewed by the UMFÍ IT Committee at least once during each board term to ensure that it is in line with the latest technology and law.
The UMFÍ Board approved the AI policy at a board meeting on 5 September 2026.