Crypto Firms Demand Stronger AI Security Access Amid Vulnerability Research Shift

Crypto companies are pushing for stronger access to AI security tools as new models change vulnerability research. The Bitcoin Policy Institute, along with over three dozen signatories including Coinbase and Block, has urged major AI labs to provide open-source security researchers with early access to their strongest cyber-capable models.

The current restrictions can leave Bitcoin defenders using weaker tools than attackers, the group says. Public models may block legitimate vulnerability research because safety filters treat some defensive security work like malicious activity.

Developers may rely on downloadable open-weight models that are less capable, while attackers use public models and specialised hacking tools without facing the same restrictions.

The request comes after BTCPay Server disclosed a critical vulnerability that had already been exploited against Lightning nodes belonging to merchants. Hardware wallet maker Foundation also lost its own node during the attack.

Meanwhile, Japanese Bitcoin treasury company Metaplanet has expanded its financing options through a continuous bond issuance program called BitBonds.