Over 40 Bitcoin firms have called on AI labs to enhance security access for researchers, highlighting a critical imbalance in cybersecurity resources.
Summary
- Over 40 bitcoin and cryptocurrency firms have signed an open letter requesting AI laboratories to implement trusted-access programs for security researchers.
- The initiative, led by the Bitcoin Policy Institute, includes prominent firms like Blockstream and ARK Invest.
- Security teams lack access to cutting-edge AI models, while attackers exploit fewer restrictions.
This week, a coalition of more than 40 bitcoin and cryptocurrency companies appealed to major AI labs to grant independent security experts access to their advanced models before they are released to the public.
The letter, dated August 10, underscores a critical issue: the Bitcoin network supports over a trillion dollars in value, with the broader ecosystem—including wallets and custody systems—representing even more. Given the public nature of this code, a single flaw could result in substantial financial losses.
However, those tasked with securing this infrastructure face challenges accessing vetted researcher programs from AI labs. Attempts to utilize public models often encounter safeguards that obstruct legitimate efforts to identify vulnerabilities. Consequently, security teams depend on open-weight systems that, while functional, do not keep pace with the latest technological innovations—leaving them vulnerable to sophisticated attackers.
The letter emphasizes, "Early access and reasonable compute budgets would enable core developers to create more secure code and address vulnerabilities proactively. We urge leading AI labs to establish or enhance trusted-access programs for qualified defenders of open-source financial infrastructure."
It also points out that AI labs and select partners often acquire new attack capabilities months ahead of the general public, with many of these tools leaking through public releases and compromised systems.
AI Labs' Response Uncertain
Recently, the Bitcoin Security Consortium—comprising firms like BlackRock, Coinbase, and Fidelity Digital Assets—allocated $15 million to enhance the long-term security of the network. However, the current letter highlights a different concern: the necessity for access to contemporary AI models for those responsible for the network's security.
As of now, major AI labs have not publicly responded to this request. Nonetheless, these organizations are actively developing improved safeguards to prevent misuse of their models, whether in blockchain initiatives or other sectors.
Anthropic has previously showcased how such models can facilitate "vibe hacking," where criminals use AI to conduct live Bitcoin ransom attacks, arguing that these tools reduce the cost of offense. Providing advanced models to defenders involves selecting trustworthy individuals and ensuring they maintain confidentiality regarding sensitive information.
The letter encourages labs to take this risk into account.
