Artificial Intelligence Policy

1% for the Planet Artificial Intelligence Policy

Last updated: July 2023

Purpose

At 1% for the Planet, we recognize the potential benefits of artificial intelligence (AI) for supporting our stakeholders, our staff, and advancing our mission. We are committed to using AI in a responsible, ethical, and effective manner, and to fostering a culture of transparency and accountability in our use of AI tools.

To achieve these goals, we have developed this AI policy to guide our use of AI and ensure that it aligns with our core values and mission. We will regularly review and update this policy to reflect changes in technology, industry standards, and legal and regulatory requirements. We subscribe to the “first draft not final draft” view of content or analyses generated by AI, and will always center our staff and stakeholders when considering the use of AI systems.

We are committed to being open to feedback and making changes as necessary to ensure that our use of AI is responsible, ethical, and mission-aligned. As the world of AI continues to evolve, we will continuously evaluate and improve upon our use of AI to ensure that it meets our high standards for ethical and responsible use.

Ethics & transparency

We will ensure that any AI systems or tools we deploy are guided by ethical principles, including an unwavering commitment to transparency in how we use AI. We will regularly review and update this publicly-available AI policy to reflect the latest ethical standards and best practices.
We will be transparent with our stakeholders about our use of AI. This includes regularly communicating with our stakeholders about the development and use of AI systems, as well as how those systems or tools make decisions and which data they use.
Any content or images generated by AI will always be reviewed and edited by human experts to ensure they are accurate, meet our ethical standards, and do not contain any objectionable or harmful content.

Images generated by AI will be accurately credited. We will respect the work of authors and artists when designing prompts to avoid directly using living artists in prompts.
Content developed with the support AI will utilize the same attribution process we would typically use for both internal and external facing content. For example, we don’t typically credit an author for a blog post, as many staff may have contributed to a given piece. When in doubt though, we will over communicate about our use of AI as an assistant in generating content.

Data privacy & security

We will use AI in ways that protect the privacy of individuals and businesses, and the security of their data.
We will ensure that any data we collect or use is handled in accordance with our existing privacy policy, best practices and applicable regulations. Sensitive or personal data will always be anonymized, encrypted, or otherwise protected if used with AI tools.

Preventing discrimination & bias

We will use AI to promote diversity, equity, and inclusion in our global network. We will work to ensure that our AI tools and services are accessible and inclusive for all.
We will prevent and address bias and discrimination in our AI systems by using diverse datasets to build any AI systems, by leaning into diverse input and expert review of AI systems, and by regularly auditing and testing for bias and discrimination.
We recognize that AI tools, and in particular Large Language Model (LLMs) are only as good as the content they are trained on, which means they are subject to the inaccuracies and biases of the training content. These LLMs can also “hallucinate” (i.e. generate factually inaccurate content that may sound correct). As such, we are fully committed to training our staff on appropriate use of LLMs and AI systems, and ensuring human oversight of any published content.

Human oversight & risk mitigation

We will consult with external experts as we develop and deploy any AI systems.
Before deploying any AI system or tool, we will conduct a risk assessment and develop a plan for mitigating and managing those risks.
We will ensure that there is human oversight of any AI systems by training staff on AI risk management, and teaching them how to monitor and regularly review those systems, and intervene when necessary.
As described above, we are committed to expert human review of any content generated with the help of AI systems, prior to sharing or publishing outside of our organization.

Regulatory compliance

We will consult with legal experts to ensure that any AI systems we develop or deploy are in compliance with applicable laws and regulations.

Training & education

We are committed to investing in AI prompt training and education for staff, including training on how to use and interpret AI systems, as well as training on the ethical considerations associated with the use of AI.
Basic AI guidelines will be required in alignment with this policy to ensure staff is trained appropriately. We also will not allow the use of prompts that mimic living creators like artists or authors.

Sustainability

We recognize the importance of minimizing the environmental impact of any AI systems we use or develop. We will account for increased energy consumption as part of our GHG Scope 3 measurement and offsetting practice

Review & updates

This policy will be reviewed annually, or as needed based on changes in technology, regulation, or organizational needs. Updates will be communicated to all users and appropriate training will be provided.