AI Policy

ClimateFast's AI Policy

Guiding principles:

1. ​​Collaboration & Community: Minimizing reliance on AI may help strengthen community building, existing information networks, and collaboration opportunities. First, we want to rely on building long-lasting relationships and skills in collaboration with others to understand complex issues. It’s tempting and easy to do a quick search using AI to unpack a complex issue, but this may be taking away from what could be learned collectively through discussion and research collaborations, especially. Among people with varied backgrounds. Second, learning about climate change through person-to-person conversations and information sharing can promote deeper learning. Solely relying on AI to understand issues may take away from understanding of an issue through debate and meaningful discussion. Lastly, we want to help sustain existing information networks and information-sharing practices with other people.
2. Building Public Trust: Using AI-generated content and search results may lead to more AI-generated content being integrated into our day-to-day organizing materials. This content can then be recycled back into the search results of AI-generated content and lead to watered-down, potentially lower-quality results. We want to maintain quality content and rely on information from primary, human-made sources whenever possible - not contribute to a potentially destructive cycle of low-quality information.
3. Energy Consciousness: We commit to minimizing the energy consumption associated with everything we do, including our AI use whenever possible. We will avoid AI for tasks that can be accomplished by a person and if need be, only employ AI tools under special circumstances when a clear and demonstrable benefit exists that cannot be achieved through non-AI means. In the absence of transparency on the energy intensity of AI, we take a cautious approach to not use it whenever possible, unless and until it is made transparent how energy and water-intensive it is, and if/how it is impacting local communities.
4. Protecting Critical Thinking: We believe in fostering human skills and critical thinking rather than diminishing them, even if it takes more time and effort to create personalized writing and art. In addition, we live in a world where there is an abundance of information and new research produced every day. It’s hard to keep up with it all. Many of the problems we see today are due to having so much 0nformation and not knowing how to process it all.
5. Bias Awareness: We acknowledge that AI models can reflect and amplify existing biases present in their training data. As we get more exposure to AI-generated images or writing, we must try to retain our critical thinking, reflect deeply, and notice biases where they occur. AI can be a black-box, provide one-sided information and a lack of clarity in how it produces its output. An example of this is the challenge of including those who are excluded from decisions and actively searching out their material (this is a part of decolonializing our activities).
6. Transparency and Accountability: We will be transparent with each other about our use of AI, including the tools we employ and the purposes for which they are used when it comes to creating content for ClimateFast. If AI is being used to create and support ClimateFast, it should be noted in documents. Moreover, the person using AI should get clarification from the ClimateFast core group in advance as to whether there is a ‘clear and demonstrable benefit that cannot be achieved through non-AI means’ (per #1 above).
7. Continuous Learning: The field of AI is rapidly evolving. We commit to staying informed about the energy intensity of AI and emerging ethical considerations. We will periodically review and revise our AI policy as new information becomes available. This also includes sharing new resources with each other on AI, pertaining to the environment and climate.

Best practices on use of generative AI at ClimateFast: 

1. The following is a list of activities at ClimateFast that will not use AI generated text or images as a best practice:

• Meeting notes will continue to be taken manually by the notetaker and not rely on AI, unless the notetaker is away and a substitute cannot be found.
• Letters to elected officials and organizations
• Social media images and captions
• Information on the ClimateFast website
• Campaign material
• Printed materials

2. If a volunteer at ClimateFast uses AI-generated research (e.g. anything saved in the CF Google Drive), writing, or images for ClimateFast purposes, the core group should be notified and the citations should be provided (source, date, medium used).

3. Where possible and relevant, ClimateFast will advocate for AI data centres to use renewable energy to power the facilities, especially the ones within Canada.

4. As individuals, we aim to take time to reflect on whether it is worth using the AI service and notify others of tips and solutions to minimize usage and/or if there are viable alternatives to using AI for the same task. For example:

• Ex. 1: Keeping each other informed about finding alternatives to AI. For example, using a regular search engine vs. searching via ChatGPT or Gemini.
• Ex. 2: Utilizing letter writing resources and templates we already have to help save time

7. The Volunteer Coordinator should provide an orientation to new recruits on our AI Policy.

8. Review the ClimateFast AI policy (including values and best practices) every six months

*The best practices above apply to ClimateFast as an organization, not to individuals in their own pursuits.