Summary
- Over 40 bitcoin and cryptocurrency companies have signed an open letter urging leading AI labs to implement trusted-access programs for qualified security researchers.
- This initiative was organized by the Bitcoin Policy Institute, with notable participants including Blockstream and ARK Invest.
- There exists a significant disparity: those tasked with securing wallets and node software lack access to advanced AI models, while attackers have no such limitations.
This week, more than 40 companies involved in bitcoin and cryptocurrency called on major AI laboratories to allow independent security experts to utilize their most advanced models prior to their public availability.
The letter, released on August 10, highlights a critical imbalance. The Bitcoin network underpins over a trillion dollars in value, while the broader ecosystem—comprising wallets, signing devices, libraries, and custody systems—accounts for trillions more. As this code is publicly available, a single vulnerability could lead to significant financial loss.
Myriad: When will OpenAI release GPT-6? Make your prediction.However, those responsible for maintaining this infrastructure cannot access the vetted researcher programs offered by AI labs. When they attempt to use public models, the safeguards often hinder their legitimate search for vulnerabilities. Consequently, they rely on open-weight systems they can operate independently, which, while useful, lag behind the latest advancements—and the hackers targeting cryptocurrency systems.
The letter states, "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. We, therefore, ask frontier AI labs to establish or expand standing trusted-access programs for qualified defenders of open-source financial infrastructure."
It also notes that laboratories and certain partners gain access to new attack capabilities months in advance of the general public, and these capabilities often leak through public releases and compromised systems.
AI Labs Remain Noncommittal
Recently, the Bitcoin Security Consortium—which includes BlackRock, Coinbase, Strategy, Anchorage, ARK, Block, Blockstream, Fidelity Digital Assets, and Galaxy—allocated $15 million for the long-term security of the network. The letter addresses a different issue: the need for access to current models for those maintaining the network's security.
As of now, prominent AI labs have not made any public statements regarding this particular request. However, these firms are actively working on improved safeguards to prevent their models from being misused, whether in blockchain development or other applications.
Anthropic has previously demonstrated how similar models facilitate "vibe hacking," where criminals conduct live Bitcoin ransom attacks with AI assistance, asserting that these tools lower the cost of offense. Providing stronger models to defenders involves selecting qualified individuals and trusting them not to disclose sensitive information.
The letter encourages the labs to embrace that risk nonetheless.