More than three dozen Bitcoin and digital asset companies, developers and cybersecurity organisations are urging frontier artificial intelligence (AI) labs to give trusted open-source security researchers access to their most capable cyber-focused models.
The open letter organised by the Bitcoin Policy Institute argues that AI is changing the balance between cyber attackers and defenders, while safety controls on publicly available models can prevent legitimate security researchers from identifying vulnerabilities. The signatories said the Bitcoin network alone secures more than $1 trillion in value, while the wider digital asset ecosystem relies on wallets, signing devices, cryptographic libraries, node software, custody systems, exchanges and payment networks.
“Across the broader digital asset ecosystem, trillions of dollars depend on” this infrastructure, the letter said.
“For millions of people, the value secured by this software is deeply personal,” it added. “It represents retirement savings, business capital, emergency funds, family wealth, and years of financial decisions.”
“A vulnerability in this infrastructure can place an entire life’s savings at risk.”
The letter said frontier AI is changing how quickly both attackers and defenders can conduct security work. “Advanced models can search large codebases, identify potential weaknesses, and accelerate complex technical work for both defenders and adversaries,” it said.
The signatories expect these capabilities to spread through public open-source models, compromised enterprise systems, and specialised cyber tools.
They warned that frontier AI labs and selected partners can see emerging cyber capabilities months before they reach the wider market, while qualified open-source defenders have limited and inconsistent access.
As a result, some researchers must rely on less capable open-weight models for critical security reviews.
“Without dedicated access programmes, defenders may lack the tools needed to keep pace with evolving threats to the infrastructure they maintain,” the signatories said.
The call comes as AI-assisted security testing is generating numerous findings across Bitcoin projects. Decrypt reported that a volunteer group called the Bitcoin Red Team filed 4,962 security findings across 390 Bitcoin projects in about 30 hours. The campaign identified 85 critical and 635 high-severity issues, with AI agents performing much of the scanning.
According to Decrypt’s report, 91 per cent of the findings came through automated scan intake, although researchers continued to guide and assess the AI systems manually. About 21 per cent had been dynamically reproduced with proof-of-concept code.
The campaign has also raised concerns about the workload facing maintainers. At the time of Decrypt’s report, only 19 projects had received upstream disclosures. The open letter said small development teams are already feeling the pressure.
The organisations are asking frontier AI labs to establish or expand standing trusted-access programmes for qualified defenders of open-source financial infrastructure. Their requests include “early, controlled access to the strongest cyber-capable models, including pre-release access where appropriate.”
They also want “sufficient compute and long-horizon agent use to conduct meaningful security reviews,” secure environments for private or embargoed code, and eligibility for small nonprofit and independent maintainers.
The letter also calls for “a direct channel for coordinated disclosure, remediation, and communication with lab security teams”.
The digital asset sector said it is prepared to help AI companies identify credible participants, establish standards and coordinate remediation.
“It is clear that early access and reasonable compute budgets would help core developers produce more secure code and identify and remediate vulnerabilities before losses occur,” the letter said.
The appeal concludes that frontier AI could become a major defensive technology if the people securing critical infrastructure can access it early.
“Frontier AI can become one of the most powerful defensive technologies ever developed,” the letter said. “Defenders need the frontier. Please give them a fair head start.”
