Rob Hamilton, CEO of AnchorWatch, announced the sudden restriction of access to cybersecurity capabilities of artificial intelligence from OpenAI. The incident occurred on August 8, 2026, when Hamilton tried to integrate these tools into his work as part of the Bitcoin Red Team volunteer initiative. Despite the fact that the specialist went through the verification procedure (KYC) in advance and successfully completed the onboarding process several months ago, access to the service was blocked the very next morning. As a result, the researcher was forced to suspend auditing changes to the software code and checking for other potential vulnerabilities, which temporarily froze some of the critical ecosystem security work, highlighting the fragility of dependence on proprietary platforms.

According to Hamilton, such a policy by American technology companies creates a dangerous imbalance in the field of cybersecurity. Restrictions cut off independent defenders and researchers from the most advanced and powerful analytical tools, while attackers operating outside the legal field do not face such barriers. Although there have been no problems with access to the company's models at the moment, the expert expressed concern about a possible downgrade of access from the Fable model to Opus, which could lead to similar restrictions. This jeopardizes the continuity of monitoring of open repositories, which requires the constant use of the most up-to-date computing power to detect threats in a timely manner.

The Bitcoin Red Team, consisting of 25 participants, specializes in using artificial intelligence and manual verification to analyze hundreds of open source repositories related to bitcoin. As of August 9, 2026, the initiative has already identified 1,288 critical and high-level vulnerabilities. Despite the fact that the OpenAI and Anthropic cyber models show "amazing results," the team was forced to return to using Chinese open AI models. Currently, the Kimi K3 model accounts for almost 75% of the group's accumulated expenses on artificial intelligence, and the share of the Qwen 3.8 model continues to grow steadily. Against the background of these events, in July, 25 American companies have already called on the US authorities to refrain from imposing strict restrictions on Chinese developments in the field of AI, warning of the risks of technological isolation.