The Bitcoin Policy Institute and dozens of organizations across the digital-asset and open-source sectors are asking leading artificial intelligence labs to give vetted security researchers controlled access to frontier AI models.
In an open letter published Monday, the coalition said developers working on digital-asset infrastructure often lack access to the most capable AI systems used for cybersecurity research. It also said safety restrictions on publicly available models can prevent legitimate security testing, leaving some researchers to use less capable open-weight alternatives.
The groups want AI labs to establish dedicated access programs that would include advanced cyber-capable models, additional computing resources and secure environments for reviewing sensitive software.
Coalition targets gap in AI security access
The letter argues that the strongest AI capabilities are increasingly concentrated among AI labs and selected partners, while many open-source maintainers do not have comparable access.
According to the coalition, this creates an imbalance for teams responsible for reviewing publicly developed software used across digital-asset infrastructure. The signatories include open-source development organizations, custodians, payment companies, security firms, investment firms and other businesses connected to the digital-asset ecosystem.
The coalition includes companies and organizations such as Coinbase, BitGo, Bitwise, Blockstream, Chaincode Labs, Exodus, Galaxy, MARA, Strategy, Strike, Trezor and Unchained.
Why AI access matters for crypto security
Digital assets rely on a wide range of open-source components, including wallet software, cryptographic libraries, blockchain nodes, custody systems and payment infrastructure.
The coalition argues that vulnerabilities in those systems can have direct financial consequences, making early identification of security flaws particularly important. It also said the Bitcoin network and broader digital-asset infrastructure collectively secure trillions of dollars, although the letter does not provide an independent methodology for that estimate.
The Bitcoin Policy Institute said it has received reports from open-source maintainers about sophisticated cyber activity involving advanced AI capabilities. Those reports were not independently detailed or verified in the letter.
BPI has pushed broader crypto policy changes
In March 2026, the Bitcoin Policy Institute also called on U.S. lawmakers to extend proposed de minimis tax exemptions to Bitcoin and other major cryptocurrencies, rather than limiting the relief to payment stablecoins.
The group argued that treating Bitcoin as property under U.S. tax rules can create a capital-gains reporting requirement for small purchases, adding record-keeping costs to everyday transactions. BPI said broader exemptions could make Bitcoin more practical for routine payments.
Groups seek controlled access, not unrestricted models
The coalition is proposing a vetting system rather than unrestricted access to frontier AI. Its requested framework would include:
- Controlled or early access to advanced cyber-capable models.
- Sufficient computing capacity for large-scale and longer-running security reviews.
- Secure environments for private or embargoed code.
- Eligibility for independent researchers, nonprofit maintainers and smaller open-source teams.
- Direct channels between researchers and AI labs for vulnerability disclosure and remediation.
The groups said the digital-asset industry could help identify qualified participants and coordinate responses when vulnerabilities are found.
Where AI safety measures meet security research
The request reflects a broader challenge for AI companies: safeguards intended to limit malicious cyber use can also restrict legitimate security research. The coalition argues that vetted programs would allow AI labs to maintain restrictions while giving trusted defenders access to capabilities needed to audit critical software.
It also wants researchers to receive access to newer systems before those tools become broadly available, arguing that defenders need time to identify and address vulnerabilities.
No AI lab commitment disclosed
The letter calls on leading AI labs to establish or expand trusted-access programs but does not identify a specific commitment from any lab to adopt the proposed framework.
For the crypto industry, the push places AI-assisted security research alongside ongoing efforts to strengthen the open-source infrastructure underpinning digital-asset networks.
Also Read: Fed Chair Kevin Warsh Completes Sale of Holdings Under Ethics Agreement
