Business Software Alliance offers federal agenda to support adoption of open-weight AI models

The Business Software Alliance makes a case in a new blog post for the federal government to help with the adoption of open-weight artificial intelligence models by providing guidance to companies, focusing on supply chain security and resilience and incorporating their use into cyber defenses. “Open-weight models — AI models whose parameters can be downloaded, run, and customized on an organization’s own infrastructure — have become foundational to improving and expanding access to AI across the economy,” BSA CEO Victoria Espinel writes in the Aug. 5 post . Espinel also recognizes that enterprise software companies and others who are building with AI are relying “on a combination of proprietary and open technologies.” She writes, “Both matter. A healthy AI ecosystem depends on both, and on organizations retaining the ability to choose between them.” Espinel’s post highlights the benefits of open-weight AI models and offers a policy approach for governments. BSA’s membership includes major software firms and tech companies. Notably, OpenAI is a member and has emphasized the importance of the U.S. government not cutting off access to open-weight models developed in China. Anthropic is not among the trade association’s membership and has taken a different approach on a potential ban. Anthropic CEO Dario Amodei in a July 27 post pushed for “requiring safety testing of all sufficiently capable models, open and closed” and encouraged “stopping industrial-scale distillation.” The BSA post looks at open-weight models through the lens of federal procurement and supporting companies with the “responsible adoption” of the models through providing “best practices for AI governance.” Espinel argues, “Give companies clear guidance that supports adoption of open models. The government is well positioned to develop practical guidance on evaluating open model provenance and verifying model integrity.” On procurement, Espinel says, “Keep federal procurement neutral. Agencies should not foreclose solutions built on open models. Federal buying practices should evaluate systems on performance, security, and fitness for purpose.” Another prong of BSA’s argument is to tie responsibility for open weight models “to conduct, not release.” Espinel explains, “What matters is how an organization deploys and governs a system — not how the underlying model was made available. The policy focus should be on supply chain security and resilience, not whether a model has open weights.” Espinel also emphasizes that “open models and tools” should be recognized as “defensive assets.” She writes, “Cybersecurity policy, including secure-by-design guidance and vulnerability disclosure programs, should reflect the role open models play in defense. Governments should also invest in shared open defensive infrastructure: evaluation frameworks, red-teaming tools, and reference datasets.” The post also comes as the White House is considering next steps for a voluntary framework that would allow the government to test frontier AI models with advanced cybersecurity capabilities 30 days ahead of their public release. The Trump administration’s June 2 executive order on frontier AI models set a Aug. 1 deadline for the departments of Treasury and Defense, the National Security Agency and CISA to develop the framework. The administration met the Aug. 1 deadline but doesn’t plan to make the framework public, according to media reports . Media reports also revealed that open weight models will be excluded from falling under the voluntary framework for testing, following an Aug. 4 meeting at the White House with AI firms to introduce the framework. Espinel argues, “Enterprise software companies build with both open-weight and proprietary models. And often the software services they provide allow their customers to bring a range of AI models – proprietary or open – to the service. Customers expect the flexibility to choose what works for them.” “The policy question is not which approach should win. It is whether organizations across the economy will retain the ability to choose — and whether governments will support responsible adoption of the full range of tools available to them,” according to Espinel. In a separate position paper , BSA argued on July 17 for C-suite executives to embrace frontier AI models with safeguards in place to rely on cybersecurity best practices. The paper says, “Frontier AI cybersecurity systems mark a structural shift in cybersecurity, which is no longer bounded by human timelines or expertise. This shift reinforces that cybersecurity has always been a core responsibility of executive leadership—and makes it more critical than ever.” -- Sara Friedman (sfriedman@iwpnews.com) Article Type: Daily News Sara Friedman tags: AI Weight: -20