The SDSN’s Core Principles on AI

The SDSN adheres to several principles across the vast array of Artificial Intelligence (AI) functions and services. The organization maintains the following principles with regard to its use of AI technologies.

Respect for Privacy: In alignment with GDPR principles, under no circumstances should individuals ever input into AI tools any confidential (e.g., donor reports, budgets) or personal information about the SDSN or anyone associated with SDSN, including staff, board members, leadership council members, volunteers, donors, or partners. Personal information comprises any data that can be used to identify someone, including but not limited to an individual’s full name, home address, email address, government-issued ID, social security number, financial information, health information, IP address, signature, etc. As an added precaution, where possible individuals should anonymize data when using AI tools. This can be done by using an anonymizer, such as MaskMyPrompt. One should also assume that anything included in an AI prompt is now public information and, when possible, should opt to not allow the AI tools to use our inputs and queries for AI training purposes.

Respecting Intellectual Property (IP) Rights of Sources: Authors and publishers own valuable intellectual property rights in their creative and research output. Individuals should not use AI tools to obtain copyrighted material unless they otherwise have permission to do so from the rightsholder. Individuals should not use AI tools to impersonate or copy the style of another author or artist. When using AI for research, one should seek sources for text and ideas presented by AI. Individuals must be diligent in thoroughly researching the outputs of AI platforms to ensure they have the right to use source material and to credit sources accordingly. The SDSN uses APA style.

Respecting IP Rights of SDSN Partners: Because many AI tools train off of user input, individuals must not input, upload, transmit, or otherwise provide any of the SDSN’s partners’ or contributors’ information, data, or IP into any AI tool without their prior written consent.

Ensuring Human Oversight, Verification, and Accountability of AI Outputs: The SDSN commits to human oversight of all AI outputs. This includes editing, review, and verification of all content (both internal and external) to prevent unoriginality, plagiarism, or other unwanted side-effects of AI influence. The SDSN takes responsibility for all material published on its platform.

Awareness of Bias: AI poses a significant risk for misinformation, and individuals need to ensure that they are not contributing to the dissemination of fallacies, discredited ideas, or ideas that are not grounded in facts. Bias may occur at any stage of the AI pipeline. To safeguard against extending unfair bias and profiling as outlined in GDPR Article 22, individuals must work through the proper data verification processes to reorient AI outputs as needed.

Transparency When Using AI: Individuals should clearly convey when Generative AI is used to assist with the creation of content.

Sustainability of AI: The SDSN acknowledges the potential harmful impact AI technology might have on sustainable development and energy systems around the globe and continues with sensitivity and awareness towards potential consequences. When considering the use of AI, individuals should first consider if there are more sustainable and energy-efficient ways to complete the task without its use.

Maintaining SDSN’s Globally Renowned Reputation and Authenticity: As a well-established science-based organization with a robust global network and brand, it is essential to the organization’s mission that we maintain the highest standards of our outputs, original thought, and creativity in the solutions and recommendations we put forward. While AI may be a powerful tool to support our mission and enhance the efficiency of some tasks, priority should be given to cultivating original ideas, writing, and creativity.

Adherence to Government-mandated AI Regulations: The SDSN commits to adhere to government-mandated AI regulations and laws. To assess AI risk, where applicable it adopts the European Union’s AI Act Framework.

*These principles were developed in April 2025 and will be reviewed and updated on a regular basis as the AI ecosystem evolves.