| Field | Value |
|---|---|
| Type | Software-industry trade association |
| Known for | Representing the business-software and enterprise-technology sector in policy debates |
The Business Software Alliance (BSA) is a software-industry trade association that represents the business-software and enterprise-technology sector in technology policy debates. It is one of several US technology trade associations, alongside CCIA, SIIA and TechNet, that intervene in federal AI policy through joint letters and comment filings.
Overview
BSA describes itself as the global trade association of the enterprise software industry, representing companies in artificial intelligence, cybersecurity, cloud computing, quantum and other technologies, and states that it works in more than 20 markets in the United States, Europe and Asia. It also promotes software license compliance and software asset management, raising awareness of the risks of unlicensed software use (Source: bsa.org). Staff who have spoken for the association on AI policy include Aaron Cooper, senior vice president for global policy, and Henry Young, senior director of policy (Source: techpost.bsa.org; Source: bsa.org).
Activities in AI policy
Frontier AI security models
In June 2, 2026 coverage, BSA proposed a voluntary, phased approach to rolling out frontier AI models with advanced cyber-vulnerability-discovery capabilities. The proposal followed the reveal of Anthropic's Claude Mythos and comparable OpenAI offerings (Source: https://insideaipolicy.com/ai-daily-news/software-group-proposes-phased-rollout-frontier-ai-models-vulnerability-discovery). The underlying proposal was a June 1, 2026 BSA TechPost post by Aaron Cooper, which named Mythos and GPT 5.5-Cyber as the models prompting the discussion. It recommended a public-private collaboration to create a voluntary phased-rollout process for models capable of substantially more sophisticated vulnerability discovery and exploitation than other available models, so that defenders could improve security before malicious actors gained access. Cooper argued that the process should be structured, transparent and globally aligned, cited Project Glasswing, OpenAI's Trusted Access for Cyber program and the Secure Future Initiative as examples of the intended objective, and pointed to the G7 Hiroshima Process and the partnership between the US Center for AI Standards and Innovation and the UK AI Security Institute as models for public-private cooperation (Source: techpost.bsa.org).
On September 9, 2026 BSA launched "Cybersecurity Professionals First," a framework for voluntary public-private partnerships giving qualified cybersecurity professionals priority access to highly capable frontier AI models. The framework recommends that government and industry identify systems for evaluating frontier AI security models, set criteria for priority access, and create mechanisms to end phased rollouts and make models broadly available. It sets out four access criteria: predictable program design (clear participation rules, evaluation process and pre-release duration); trust and transparency (a history of responsible conduct and compliance, the ability to limit unintended access, and transparency of ownership and control); capability and capacity (the expertise and operational maturity to use frontier models for vulnerability disclosure and remediation); and mission and reach (established defensive-cybersecurity actors positioned to deliver public benefit) (Source: bsa.org).
Earlier, BSA co-signed with TechNet a May 13, 2026 open letter led by the Independent Community Bankers of America to the Office of the National Cyber Director, which urged "sustained collaboration among frontier AI developers, relevant Federal agencies, and critical infrastructure stakeholders" on testing, evaluation and red-teaming of advanced models, and which favored voluntary coordination over mandatory pre-release model review (Analysis: Lawmakers, industry pitch frontier AI governance approaches as they await White House moves (Inside AI Policy, May 15 2026); see AI Pre-Release Vetting).
Federal procurement and the Anthropic dispute
BSA joined CCIA, SIIA and TechNet in a March 4, 2026 joint letter to the White House, "Letter to the President: Advancing American AI Innovation," concerning the administration's dispute with Anthropic and what the associations described as risks to US technology leadership (Source: ccianet.org). As the comment deadline on the General Services Administration's proposed AI contract clause approached in July 2026, BSA requested nine changes to the clause, including scoping revisions (Source: insideaipolicy.com; see GSA — General Services Administration (AI Deployer)).
State legislation
BSA's opposition to what coverage dated May 5, 2026 called a deployer-friendly liability regime, in a Democratic proposal to replace the Colorado AI Act, became public that week (Source: insideaipolicy.com). The article body was not read; the bill number and the specific grounds of BSA's objection are not stated in the available text.
Earlier federal consultations
BSA, listed as "BSA | The Software Alliance," submitted a response to the Office of Science and Technology Policy's request for information on biometric technologies and took part in OSTP's private-sector and civil-society stakeholder meetings during development of the October 2022 Blueprint for an AI Bill of Rights (OSTP Blueprint for an AI Bill of Rights (October 2022)).
Relationships
- related: Claude Mythos Preview, Anthropic, OpenAI
- related: Autonomous cyber-agents, AI and Cybersecurity
- related: CCIA — Computer & Communications Industry Association, Software & Information Industry Association (SIIA) — peer trade associations on the March 2026 joint letter
- related: AI Pre-Release Vetting, Analysis: Lawmakers, industry pitch frontier AI governance approaches as they await White House moves (Inside AI Policy, May 15 2026) — May 2026 ICBA-led coalition letter
- related: GSA — General Services Administration (AI Deployer) — commenter on the GSA AI contract clause
- related: Colorado AI Act (SB 24-205), AI Industry Lobbying — The 2025-2026 Political Offensive
- related: Project Glasswing: Securing Critical Software for the AI Era, NIST CAISI (Center for AI Standards and Innovation)
Membership lists and founding details are omitted pending a primary source. Confidence is rated medium: the frontier-model proposals rest on BSA's own publications, while the procurement, Colorado and coalition-letter activity rests on single news-cycle or letter sources.