The Horizon Institute for Public Service, a nonprofit organization with extensive ties to the tech-adjacent effective altruism movement, has deployed a small army of fellows at influential think tanks and government offices across Washington, D.C. At times, these fellows have worked on policy affecting the goals of effective altruists.
Effective altruists broadly believe that the development of artificial intelligence, if not carefully managed, could have apocalyptic consequences. As an outgrowth of this belief, many in the movement support export controls on cutting-edge chips to China, given the country’s focus on military applications for Artificial Intelligence technology and its lack of research guardrails.
Indeed, two profiles the Horizon Institute published on its fellows explicitly mention export controls. One fellow, who was placed at the House Foreign Affairs Committee, is said to have done work “supporting export controls” during her time in the fellowship. The other, a data scientist by trade who completed his fellowship at the Bureau of Industry and Security, researched “export controls and their ability to mitigate converged AI and bio risks” during his tenure.
Both received offers to continue their work in government following the fellowship, highlighting how the Horizon Institute’s influence can persist by translating temporary fellowships into full-fledged employment.
More recently, Sen. Jim Banks (R-IN), who hosted a Horizon Institute fellow in his office, introduced the GAIN AI Act in August, which aims to give United States customers priority over foreign businesses in purchasing top-of-the-line chips by amending the Export Control Reform Act of 2018. Legislators have since folded the bill into the Senate’s version of the National Defense Authorization Act.
Banks’s office did not respond when asked if its Horizon Institute fellow played a role in shaping the legislation.
Prior to drafting the GAIN AI Act, and while hosting its Horizon Institute fellow, Banks’s office drafted letters alongside Sen. Elizabeth Warren’s (D-MA) office raising concerns about chipmaker NVIDIA’s relationship with China, especially its alleged role in exploiting gaps in export control regulations.
A source close to the institute told the Washington Examiner that, based on their experience socializing with the fellows, they possess a wide range of views related to export controls and effective altruism, with some being skeptical of the ideas. Some fellows have published essays criticizing the apocalyptic arguments made by some effective altruists regarding AI and advocating against overly burdensome regulation of AI.
In its description, the Horizon Institute states that its fellowship program aims to train the “next generation of emerging technology policy talent” to effectively develop and implement policies related to the technology sector, while also helping to meet the demand from existing policymakers for tech-savvy staff.
“Horizon helps policymakers access deeply needed technical talent, and we do not take or advance particular policy positions,” the organization told the Washington Examiner. “Our fellows span the ideological spectrum, as do our many donors, and fellows’ work is entirely driven by their host organization.”
National security hawks and effective altruists tend to argue that export controls are an important tool for mitigating the risk of China leveraging AI to expand its military capabilities while ensuring that America has access to the necessary hardware to support its tech sector. Others, including OpenAI CEO Sam Altman, however, have argued that American firms are not currently constrained by a lack of access to high-end chips.
The Horizon Institute’s links to effective altruism run deep. The organization received its initial $2.9 million in seed funding from Open Philanthropy, a nonprofit organization run by Meta co-founder Dustin Moskovitz, who has publicly identified with effective altruism, and his wife. Open Philanthropy has been described as a “pillar” of the effective altruism community, funding a range of other groups aligned with its cause.
Good Ventures, another nonprofit organization run by Moskovitz and his wife, has also been a major supporter of the Horizon Institute, donating $11.8 million in 2023.
The Silicon Valley Community Foundation, a donor-advised fund that serves as a pass-through funder for many of the wealthiest individuals in the tech world, has allocated millions more dollars to the Horizon Institute since its inception.
Beyond funding, the Horizon Institute’s co-founder and its senior director for programs both have ties to effective altruism through their previous work.
The duo previously worked at the Center for Security and Emerging Technology, a think tank based at Georgetown University and founded by Jason Matheny, the CEO of the RAND Corporation. Matheny has been described as a “luminary in the effective altruist community.”
RAND Corporation significantly influenced the Biden administration’s AI policy, with Matheny claiming at an October 2023 internal meeting that analyses produced by his organization “informed new export controls and a key executive action expected from the White House next week.”
Matheny also sat on the Horizon Institute’s advisory board at one point.
Tim Stretton, director of the congressional oversight initiative at the Project on Government Oversight, has criticized the Horizon Institute, arguing that it is improper for its fellows to be working on policies that their funding organization has an interest in.
“We have [the] AI [industry] inserting its staffers into Congress to potentially write new laws and regulations around this emerging field,” Stretton said. “That is a conflict of interest.”
The Horizon Institute, however, has defended its fellowship program by stating that it is “not for the pursuit of particular policy goals” and that it does not screen prospective fellows for ideological alignment.
MEET THE BILLIONAIRE FUNDING CONSENSUS AT EVERY LEVEL OF THE HEALTHCARE DEBATE
Nonprofit organizations, such as the Horizon Institute, can, under the Intergovernmental Personnel Act of 1970, pay the salaries of fellows working in Congress or in the president’s administration. Through this arrangement, the Horizon Institute has placed nine fellows in congressional offices held by Republicans and six fellows in congressional offices held by Democrats, as well as the Department of Defense, Department of Commerce, and the White House. Premier think tanks, such as the Heritage Foundation and the RAND Corporation, which play significant roles in developing AI policy, have also recruited fellows from the Horizon Institute.
The influence of these fellows doesn’t necessarily wane after the term of their fellowship expires, as many receive full-time job offers from the offices they worked in or in another government role.