Responsible AI Policy

C-CAS recognizes the importance of responsible use and development of artificial intelligence (AI) in chemistry. While the tools created and used by C-CAS have the potential to benefit the field of chemistry and society at large, the Center also recognizes that there are risks associated with potential misuse of these tools, as well as potential costs. To encourage thoughtful reflection on the responsible use of AI, a series of questions (shown below) are included in the biannual reports required for all C-CAS projects. The goal of this evaluation is to encourage C-CAS members to consider the potential impact of their work and to implement strategies to minimize or mitigate potential costs.

• Who are the intended users of this work?

• Are there any concerns regarding ownership of data used in this tool or model?

• What are the potential benefits to society (human health, environment, etc.) if this project is successful?

• What are potential risks of intentional misuse of this technology? How do you plan to minimize or mitigate these risks?

• What are potential risks of accidental misuse of this technology? How do you plan to minimize or mitigate these risks?

• What are the costs of pursuing such a project (e.g., use of large-scale compute with environmental impacts, material/waste requirements)? How do you plan to minimize or mitigate these costs?

Answers to these questions are reviewed centrally by the Center Director and staff. Any areas of concern are brought to the C-CAS Responsible AI faculty working group which evaluates whether concerns require additional intervention, such as a) stopping work, b) rescoping, or c) refraining from releasing a model openly. If the RAI working group suggests that action is needed, the issue is brought to the C-CAS Executive Committee to approve and implement any intervention.