Inside OpenAI’s Plan to Make AI More ‘Democratic’

Posted under: AI technologies
Date: 2024-02-08
Inside OpenAI’s Plan to Make AI More ‘Democratic’

In May 2023, OpenAI sought Colin Megill's help to align AI with human values, inspired by his work on Polis. They proposed using large language models (LLMs) to facilitate deliberations. Despite initial enthusiasm, OpenAI faced turmoil as its CEO was briefly ousted. Despite skepticism, Andrew Konya's project with Remesh received funding. While promising, it grappled with the limitations of LLMs and the challenge of democratic representation. OpenAI's program aimed to solicit public input on AI governance, but its binding nature remained uncertain. Despite this, OpenAI's efforts have invigorated deliberative technologies, with hopes for public ownership and influence over AI development. Yet, questions persist about the true extent of public power in shaping AI's trajectory amidst competitive pressures and corporate interests.

Read more at: time.com