Could AI Speak on Behalf of Future Humans?

In a recent piece in the Stanford Social Innovation Review, visiting researcher Konstantin Scheuermann and Dr. Angela Aristidou explore how Generative AI models can serve as a voice for traditionally marginalized stakeholders, such as the environment and future generations, thereby enriching collective decision-making with diverse perspectives.

A persistent challenge in many decision-making scenarios is the negligent or intentional exclusion of specific stakeholders, resulting in a perspective deficit. This exclusion stems from the biases and motivations of the individuals steering these processes.

Scheuermann and Aristidou introduce the concept of “AI Voice,” a term they coin to describe the role of AI in articulating the concerns and perspectives of these silent stakeholders (e.g., nature, future generations, animals). AI Voice encompasses the generation of human-friendly outputs by AI systems, allowing these systems to act as proxies for those otherwise overlooked in decision-making dialogs. The authors expand on the utility of AI Voice, outlining four key roles AI can play in decision-making: 1. Facilitator, 2. Consultant, 3. Optimizer, 4. Collaborator.

Acknowledging the inherent biases present in AI models, Scheuermann and Aristidou argue that these limitations can be mitigated through expert involvement in AI model development, a commitment to transparency in AI application, and efforts to improve AI and data literacy among the public.