An influential think tank, the RAND Corporation, has come under scrutiny for its involvement in shaping the United States’ AI policy. A recent revelation suggests that RAND significantly drafted President Joe Biden’s executive order on artificial intelligence (AI), specifically focusing on reporting requirements for advanced AI systems. While this move is intended to mitigate the potentially catastrophic risks associated with AI, concerns have emerged regarding the influence of Open Philanthropy, a major donor to RAND.
RAND’s influence on AI policy
President Biden’s executive order on AI, issued in October, outlines reporting requirements to regulate the use and development of advanced AI technologies. These requirements, which resemble the policy priorities of Open Philanthropy, seek to address the existential risks posed by AI, such as the potential creation of bioweapons using AI systems.
Notably, Jason Matheny, RAND’s CEO, and Jeff Alstott, a senior information scientist at RAND, are prominent figures in the effective altruism movement, which advocates for data-driven approaches to philanthropy and emphasizes the importance of addressing AI’s apocalyptic potential.
Open philanthropy’s contribution
Open Philanthropy, funded by Facebook co-founder Dustin Moskovitz and his wife Cari Tuna, is a significant contributor to RAND, having donated over $15 million this year alone. The organization aligns its philanthropic efforts with effective altruism, which places AI technology at the forefront of its concerns.
Effective altruists, including those at RAND, advocate for the responsible development and regulation of AI to prevent the misuse of this powerful technology. Their influence has grown in Washington as policymakers increasingly acknowledge the need to address AI’s potential dangers.
Critics argue that the intense focus of effective altruism on AI may inadvertently serve the interests of tech giants by diverting policymakers’ attention from existing AI-related issues. Such issues include the perpetuation of racial bias in AI systems and the erosion of copyright protections.
This criticism raises questions about the potential bias in AI policy development. RAND personnel with ties to effective altruism participate in drafting executive orders and influencing the direction of AI regulation.
In response to these concerns, RAND spokesperson Jeffrey Hiday has acknowledged that RAND personnel played a role in drafting the AI executive order. He explained that RAND’s core mission is to research and analyze critical topics and share that expertise with policymakers. However, Hiday did not directly address concerns about the organization’s objectivity in light of its ties to Open Philanthropy.
The recent revelations about RAND’s involvement in shaping AI policy have not gone unnoticed among its employees. During an all-hands meeting in October, one employee expressed concerns that the think tank’s connection with Open Philanthropy might undermine its reputation for rigor and objectivity. There are worries that RAND’s association with effective altruism could lead to prioritizing the “effective altruism agenda” over impartial policy development.
RAND Corporation’s substantial influence in shaping President Biden’s AI policy has raised questions about the organization’s objectivity and potential bias. While the think tank asserts that its involvement aligns with its mission to provide expertise to policymakers, concerns persist about the influence of Open Philanthropy and the effective altruism movement on AI policy development.
As AI continues to advance and play an increasingly pivotal role in society, it is essential for policymakers, think tanks, and organizations involved in shaping AI policy to maintain transparency and impartiality to ensure this transformative technology’s responsible development and regulation. The debate over how to strike this balance between innovation and safeguarding against potential risks will likely continue as AI technology evolves.