OpenAI Launches Program to Award $100,000 Grants for AI Governance Experiments

OpenAI, the company behind the widely-used ChatGPT, has announced the launch of a grant program aimed at fostering innovative approaches to AI governance.

Through this initiative, OpenAI aims to award 10 grants, each worth $100,000, to teams from around the world to develop proofs-of-concept for a democratic process that can shape the behavior and rules of AI systems.

This move is likely a response to the mounting concerns of AI as governments around the world race to craft regulations to govern this ever-growing technology, with issues of bias, privacy, and misinformation at the forefront of these concerns.

This also coincides with OpenAI's CEO Sam Altman traveling around the world to meet with government leaders and address their concerns about AI.

The program seeks to explore key questions and develop novel tools that can contribute to the future governance of AI.

All About the Grant Program

In a statement, OpenAI emphasized the significance of AI's impact on various aspects of society and stressed the importance of incorporating diverse perspectives in decision-making processes surrounding AI behavior.

The organization expressed its belief that decisions about AI conduct should not be dictated solely by individuals, companies, or nations. Instead, OpenAI advocates for democratic methods that consider the public interest and aim for inclusivity in shaping the future of AI.

The grant program encourages individuals, teams, and organizations to submit applications outlining their proposed approaches to address specific policy questions.

The application period will close on June 24, 2023, and successful applicants will be notified by July 14, 2023.

Selected recipients will receive a $100,000 grant to implement their proposed proof-of-concept, involving at least 500 participants, and publish a public report on their findings by October 20, 2023.

The results of the studies conducted by the grant recipients will be freely accessible, and any code or intellectual property developed during the project will be required to be made publicly available under an open-source license.

Policy Questions

The policy questions under consideration for the grant program cover a wide range of topics, including model behavior, AI assistants' responses to public figure viewpoints, the provision of medical, financial, or legal advice by AI assistants, and the balance between personalization and boundaries in AI assistants aligning with user preferences.

Other questions explore the handling of sensitive information by AI models, including the identification of gender, race, and emotions from images, as well as the depiction of individuals in generative model outputs.

OpenAI acknowledges that while the specific answers to these questions are important, the primary objective of the grant program is to foster innovation in democratic processes for governing AI behavior. The organization seeks advancements in the process itself, recognizing the complexity and wide-ranging impact of AI technologies.

To learn more about the comprehensive questions provided by OpenAI, click here.