The Bitcoin Policy Institute, joined by a broad coalition from across the digital-asset industry, published an open letter on August 10 calling on the world's leading AI labs to give qualified open-source security maintainers trusted access to frontier AI models. The letter asks labs to establish dedicated access programs that include advanced cyber-capable models, additional computing resources, and secure environments for reviewing sensitive software, according to reporting on the letter.
The coalition's core argument is one of asymmetry: the most capable AI systems are increasingly concentrated in the hands of the labs that built them and a small circle of selected partners, while the volunteer maintainers who review publicly developed, widely used open-source software have no comparable access to those same tools.
A response to a live threat
The letter isn't a theoretical policy ask. The Bitcoin Policy Institute said it has received reports from open-source maintainers describing sophisticated cyber activity involving advanced AI capabilities, and the timing lines up with a concrete recent incident: Boltz, a non-custodial Bitcoin swap provider, shut down its service in early August after a surge of AI-assisted exploit attempts outpaced its small team's ability to patch vulnerabilities. Boltz confirmed no user funds were lost, but the episode became an immediate reference point for how quickly automated, AI-driven attacks can overwhelm the security resources of smaller open-source infrastructure teams.
Because so much of bitcoin and broader crypto infrastructure, wallets, node software, Lightning implementations, bridge and swap protocols, is maintained by small open-source teams rather than well-resourced corporations, the letter frames the access gap as a direct financial-security issue rather than an abstract fairness concern. Vulnerabilities in that software can translate immediately into stolen funds, and the letter's signatories argue that defenders should not be structurally outgunned by attackers who already have access to the same frontier models.
Related: Blockstream Launches Bitcoin Swaps as Boltz Exits Over AI Attacks
What the coalition is asking for
Specifically, the letter calls for AI labs to grant vetted open-source maintainers access to advanced, cyber-capable models comparable to what commercial security partners already receive, along with the compute resources and secure review environments needed to actually use that access for code auditing. It does not name individual AI labs as targets, framing the request instead as an industry-wide standard that major labs should adopt collectively.
Whether frontier labs respond with a formal access program remains to be seen. But the letter adds digital-asset advocacy groups to a growing list of industries, alongside cybersecurity firms and critical infrastructure operators, publicly arguing that the AI safety conversation needs to account for defenders as well as the capabilities being built, particularly as incidents like the Boltz shutdown demonstrate how quickly the balance of power between attackers and small defenders can shift.
