The Bitcoin Policy Institute (BPI) has issued an open letter, dated August 10, 2026, calling on leading artificial intelligence labs to give vetted open-source security researchers and Bitcoin Core developers early access to their most advanced AI models. The request comes as AI-driven attacks on small open-source maintenance teams increase, and cutting-edge models become capable of scanning large code bases, spotting subtle weaknesses, and accelerating complex security analysis.
More than 40 organizations across the digital asset sector joined the initiative, including Coinbase, Block, Strategy, MARA, Galaxy, BitGo, Chaincode Labs, OpenSats, and Trezor. The letter argues that open-source code underpins financial systems holding well over a trillion dollars in Bitcoin alone, and many assets function as bearer instruments, meaning a successful breach can produce irreversible financial loss. Yet open-source maintainers often lack entry into internal cyber-research programs and face safety restrictions that block legitimate vulnerability hunting on frontier tools.
The coalition asks AI laboratories to create or expand trusted-access programs with pre-release model access where suitable, adequate computing resources, secure settings for sensitive or embargoed code, eligibility for independent maintainers and nonprofits, and direct communication with lab security teams. BPI reports that volunteer efforts using currently available models have already surfaced thousands of potential issues across hundreds of Bitcoin-related projects, while sophisticated actors—potentially state-linked—are already using advanced AI to sustain attacks. The central goal is to give defenders a genuine head start rather than leaving them perpetually behind adversaries.
