Bitcoin Policy Institute and more than 40 digital-asset organizations have called on leading AI laboratories to give qualified open-source security researchers controlled access to frontier models as AI-assisted cyberattacks grow more capable.
Bitcoin coalition asks AI labs for controlled access
Bitcoin Policy Institute announced the initiative in an Aug. 10 X post, saying the coalition represents organizations from across the digital-asset ecosystem.
Today, Bitcoin Policy Institute and a broad coalition from across the digital-asset ecosystem are publishing an open letter calling on the world’s leading AI labs to provide qualified open-source defenders with trusted access to frontier AI models.
The past several weeks have… pic.twitter.com/4C1tjvGiCe
— Bitcoin Policy Institute (@bitcoinpolicy) August 10, 2026
The open letter asks leading AI developers to provide qualified open-source defenders with trusted access to their most capable models. It does not call for unrestricted public access to models with advanced cybersecurity capabilities.
Instead, the signatories proposed a controlled program covering early access to frontier cybersecurity models, sufficient computing capacity, secure research environments, and direct channels with AI laboratory security teams.
Those resources would allow vetted researchers to examine Bitcoin wallets, payment infrastructure, and other open-source software before attackers can exploit newly discovered weaknesses.
“The past several weeks have made the need for this abundantly clear,” Bitcoin Policy Institute wrote.
Block, Coinbase, Strategy, MARA, Galaxy, BitGo, Brink, OpenSats, Chaincode Labs, Spiral, Trezor, Unchained, Btrust, and Fedi were among the organizations supporting the request.
The coalition argued that current model safeguards can restrict legitimate security research even as criminals and state-backed groups gain access to increasingly capable open or locally deployed systems. Its central message to AI laboratories was that defenders need an early opportunity to identify and repair vulnerabilities.
No leading AI laboratory had publicly announced a program responding to the coalition’s specific requests at the time of writing.
AI-assisted attacks raise pressure on crypto firms
The appeal follows mounting evidence that hackers are incorporating artificial intelligence into attacks against cryptocurrency companies and financial institutions.
North Korea-linked hacking group Kimsuky has reportedly built three local AI environments using Ollama, GPT4All, and Msty. The systems can support malware development, data analysis, phishing campaigns, and attack automation without transmitting sensitive queries to an external provider.
As crypto.news previously reported, the group produced AI-generated phishing material aimed at cryptocurrency, investment, and financial technology companies.
Local models create a particular challenge for model providers. An attacker running an open model on privately controlled hardware may not be subject to the monitoring, usage restrictions, or account suspensions imposed by commercial AI services.
That difference sits at the center of the coalition’s argument. Restricting trusted defenders may offer limited protection if malicious actors can use locally deployed models or bypass commercial safeguards while researchers remain unable to test the most capable systems.
The threat also carries a U.S. security angle. Several signatories, including Coinbase, Strategy, Block, MARA, and Galaxy, are publicly traded or U.S.-based companies with substantial exposure to Bitcoin infrastructure. Successful attacks could affect American customers, institutional custodians, and investors even when the vulnerable software is maintained by a global open-source community.
Bitcoin flaws show potential role for frontier models
Recent Bitcoin security incidents have demonstrated how weaknesses in open-source software can remain undetected for years.
A firmware build error affecting Coldcard hardware wallets reportedly weakened the entropy used to generate seed phrases. Attackers could then search a much smaller range of possible keys and drain wallets without obtaining physical access to the devices.
Galaxy Research estimated that confirmed Coldcard thefts reached 1,596 $BTC, while a suspected additional attack wave could raise losses to about 2,055 $BTC. The error had reportedly been present since 2021.
AI-assisted analysis subsequently expanded scrutiny beyond the original flaw. A crypto.news investigation found that automated reviews were surfacing similar classes of security weaknesses across the broader Bitcoin ecosystem.
Bitcoin Red Team provided another example of defensive AI use. The volunteer initiative reported finding 4,962 potential issues across 390 Bitcoin-related projects during fewer than 30 hours of AI-assisted code reviews.
Of those findings, 720 were initially classified as high or critical severity, although automated findings still require human reproduction and verification. More than one-fifth had reportedly been reproduced when crypto.news covered the review campaign.
Those results illustrate both the potential and limits of frontier models. AI can examine large codebases faster than small volunteer teams, but qualified researchers are still needed to remove false positives, test exploitability, and disclose verified vulnerabilities safely.
Bitcoin security funding expands beyond AI access
The open letter adds to a broader push to increase resources available to Bitcoin security researchers.
Strategy, BlackRock, Coinbase, and six other companies recently established the Bitcoin Security Consortium. Its members pledged $15 million over three years to fund developers and researchers working on Bitcoin’s long-term security.
Post-quantum cryptography is the consortium’s initial focus, though its mandate covers broader protocol security. Members said they would choose funding recipients independently and would not direct Bitcoin development or take positions on proposed protocol changes.
Galaxy separately opened a $5 million Bitcoin security fund covering new signature systems, wallet-migration tools, audits, and research into quantum-resistant protections.
Industry-wide losses continue to increase despite those initiatives. Immunefi reported that crypto projects lost roughly $110 million to hacks in July. The security platform recorded 164 incidents through Aug. 3 and projected that the number of hacks exceeding $1 million could reach a record 114 in 2026.
The Bitcoin Policy Institute coalition is now asking AI companies to complement financial support with technical access. Whether laboratories accept that proposal will depend on their ability to verify researchers, supervise potentially sensitive work, and prevent advanced cybersecurity models from being redirected toward offensive activity.
