TLDR
- Cathie Wood backed calls for qualified Bitcoin security researchers to access frontier AI models.
- Bitcoin Policy Institute wants trusted AI access programs for open-source financial infrastructure defenders.
- The proposal seeks controlled access to advanced cyber-capable AI models for qualified researchers.
- Coalition wants secure AI environments for reviewing private, unreleased and open-source code.
- Proposal covers Bitcoin wallets, custody systems, exchanges, payment networks and cryptographic libraries.
ARK Invest CEO Cathie Wood has backed an open letter from the Bitcoin Policy Institute calling on leading artificial intelligence developers to give qualified open-source security researchers access to advanced AI models. The proposal focuses on protecting Bitcoin and other digital asset infrastructure as AI becomes more widely used in cybersecurity.
Wood said attackers are already using AI and argued that security teams need access to comparable technology. “AI could become the most powerful defensive technology in history,” she said. Wood described Bitcoin as critical public financial infrastructure and said those responsible for securing it should have trusted access to advanced AI systems.
Bitcoin Policy Institute Calls for Trusted AI Access
The Bitcoin Policy Institute published the letter with support from a coalition of organizations across the digital asset sector. The group includes open-source development organizations, crypto custodians, treasury companies, payment providers, security firms and capital allocators.
The coalition wants frontier AI developers to create dedicated access programs for qualified researchers protecting open-source financial infrastructure. Such programs would provide controlled access to advanced cyber-capable models, including pre-release systems when appropriate.
The proposal also calls for enough computing capacity to conduct detailed security reviews and secure environments where researchers can examine private or unreleased code. Small nonprofit groups and independent developers would be eligible alongside larger companies and government organizations.
A direct communication channel between researchers and AI security teams is another part of the proposal. Such a system would support coordinated disclosure when researchers discover software vulnerabilities requiring fixes.
Bitcoin Developers Face AI-Enabled Cyber Threats
The Bitcoin Policy Institute said some open-source developers currently have limited access to the strongest AI cybersecurity models. Public AI services can also restrict certain security-related requests through their safeguards, leaving researchers to use open-weight models with different capabilities.
“Many digital asset defenders, including Bitcoin Core developers, lack access to lab cyber access programs,” the letter stated. The coalition said this access gap can limit legitimate reviews of wallets, cryptographic libraries, protocols and other open-source systems.
The institute also said it has received reports from open-source maintainers describing sophisticated actors using advanced AI capabilities during cyber activity. Some reports cited possible foreign adversaries, although the institute did not publicly identify the actors or provide evidence linking specific incidents to governments.
Bitcoin developers and other researchers already use available open-weight AI models to examine code and identify vulnerabilities. The coalition is seeking earlier access to more capable systems so researchers can conduct security reviews before similar AI capabilities become widely available.
Coalition Seeks AI Tools for Digital Asset Security
The letter covers infrastructure beyond Bitcoin Core. Wallets, signing devices, custody platforms, exchanges, payment networks and cryptographic libraries all form part of the digital asset systems the coalition wants qualified researchers to protect.
The Bitcoin Policy Institute said the digital asset industry could help AI companies identify eligible researchers and establish access standards. Industry participants could also assist with coordinated responses when researchers identify software vulnerabilities.
Wood supported the proposal by saying “defenders cannot be left behind” as AI capabilities advance. The letter does not seek unrestricted public access to frontier cyber models, instead calling for controlled programs that provide vetted open-source defenders with access under defined security conditions.
