The R Street Institute has proposed a set of recommendations for policymakers navigating the ongoing debate over open-source versus closed-source artificial intelligence. The organization suggests that instead of adopting “all-or-nothing” approaches, federal guidelines should be established to provide best practices for deploying open-source AI models. The think tank also calls for partnerships between government and the private sector to develop validation methods and implement “risk-tiered” liability protection.
“Rather, the path forward lies in crafting flexible solutions that mitigate the challenges and potential risks of open-source AI while unlocking its capacity to accelerate innovation at unprecedented speed and scale,” wrote Haiman Wong, resident fellow at R Street Institute, in her analysis.
Wong outlined five policy steps aimed at promoting secure development and deployment of open-source AI. She stated, “Collectively, these recommendations chart a balanced path toward securing open-source innovation — not only as an immediate national security imperative, but as a strategic foundation for sustained U.S. leadership in emerging technological domains like AI agents and robotics.”
The discussion comes as President Trump’s AI action plan addresses issues surrounding “Open-Source and Open-Weight AI.” The plan asserts that decisions about whether to use open or closed models should remain with developers but encourages the federal government to foster a supportive environment for open models. One recommendation includes ensuring access to large-scale computing power for startups and academics by improving financial markets related to computing resources.
Stanford University’s Institute for Human-Centered AI described the action plan as “the strongest federal endorsement to date of open-source and open-weight AI models.” However, Stanford researchers noted that while the plan emphasizes technical evaluations and information gathering—a move they see as supporting evidence-based policymaking—it does not fully address key risks associated with the technology. They said, “The plan’s emphasis on technical evaluations and other information gathering mechanisms reflects an important move toward evidence-based policymaking. However, it leaves key risks underaddressed.”
Wong observed that proprietary models continue to dominate U.S. AI development but pointed out that private-sector initiatives such as Meta’s Llama models and OpenAI’s anticipated open-source model have contributed to growing momentum behind open-source AI. She argued that policymakers could leverage this progress by making open-source AI a national security priority.
“If left unchecked and unchallenged, China’s open-source AI initiatives could erode U.S. technological leadership and threaten national security,” Wong wrote.
