In a coordinated effort that underscores the growing tension between artificial‑intelligence safety protocols and the practical needs of cybersecurity professionals, a coalition of leading cryptocurrency companies has formally appealed to major AI research labs. The signatories—among them Coinbase, Block, BitGo, and dozens of additional firms operating within the digital‑asset ecosystem—have authored a public letter that argues the current safety guardrails imposed on frontier AI models are inadvertently hampering essential defensive work. At the same time, they contend that malicious actors continue to exploit these same technologies without encountering comparable limitations, creating an uneven playing field that threatens the security of the entire blockchain sector. The letter, which was circulated to the press and posted on the companies’ joint website, outlines a clear set of concerns.
First, the authors point out that many of the most advanced language models and generative AI systems are equipped with built‑in safeguards designed to prevent the generation of harmful or illicit content. While these measures are well‑intentioned—aimed at curbing the spread of disinformation, hate speech, and instructions for wrongdoing—they also restrict legitimate security researchers from accessing the very capabilities they need to test and harden their systems. For example, a security analyst attempting to simulate a phishing attack, craft a sophisticated social‑engineering scenario, or reverse‑engineer a smart‑contract vulnerability may find the AI refusing to comply with seemingly innocuous prompts. This friction slows down the discovery of critical flaws and delays the deployment of patches or mitigations.
Conversely, the letter emphasizes that threat actors operating outside the law are not bound by the same ethical or policy constraints. Bad actors can simply bypass the guardrails by using open‑source models, employing private instances of AI services that lack robust oversight, or even fine‑tuning existing models on illicit data sets. In practice, this means that while defenders are forced to work around cumbersome restrictions, attackers can freely generate phishing emails, develop obfuscation techniques, or automate the creation of malicious code with relative ease.
The asymmetry, the signatories argue, undermines the overall resilience of the cryptocurrency infrastructure, which already faces a high incidence of hacks, ransomware, and fraudulent schemes. To illustrate the practical impact, the letter cites several recent incidents where AI‑enhanced tools played a pivotal role. In one case, a group of hackers leveraged a large language model to draft convincing phishing messages that targeted high‑value wallet custodians, resulting in the theft of millions of dollars worth of digital assets. In another instance, malicious actors used AI‑generated code to automate the discovery of vulnerabilities in smart‑contract platforms, allowing them to exploit those weaknesses before the developers could issue fixes.
These examples serve to highlight how the unrestricted availability of advanced AI capabilities can accelerate the speed and sophistication of attacks. The coalition’s request to AI labs is multi‑pronged. Primarily, they ask for a more nuanced approach to safety mechanisms—one that differentiates between benign research and malicious intent. They propose the creation of a vetted, credentialed access program that would grant approved security teams the ability to query advanced models without triggering blanket refusals.
Such a program could involve background checks, contractual obligations, and audit trails to ensure that the information is used responsibly. Additionally, the companies suggest the establishment of a transparent appeal process, allowing researchers to contest refusals and receive explanations for why certain prompts were blocked.
Beyond procedural changes, the letter calls for increased collaboration between AI developers and the cybersecurity community. By forming advisory boards that include experts from the crypto sector, AI labs could gain insight into the specific threats facing decentralized finance and tailor their safety policies accordingly. This partnership could also facilitate the sharing of threat intelligence, enabling AI models to be trained on real‑world attack patterns in a controlled environment, thereby improving their ability to detect and neutralize malicious behavior. The signatories acknowledge that there are genuine risks associated with loosening restrictions, but they argue that a balanced solution is both possible and necessary.
They point out that many industries—such as medicine, aviation, and finance—already operate under regulated frameworks that allow professionals to access powerful tools while maintaining strict oversight. By adopting a similar model for AI, the technology could be harnessed to bolster security rather than inadvertently weakening it. Industry reactions to the letter have been mixed. Some AI researchers praise the initiative as a constructive step toward aligning safety protocols with real‑world security needs.
Others caution that any relaxation of guardrails must be accompanied by rigorous monitoring to prevent accidental leaks of dangerous capabilities. Nonetheless, the core message resonates: without equitable access to advanced AI tools, defenders risk falling behind a rapidly evolving adversarial landscape. For the cryptocurrency community, the stakes are particularly high. Blockchain networks rely on cryptographic integrity, consensus mechanisms, and decentralized governance—all of which can be targeted by AI‑driven attacks.
As the sector continues to mature and attract mainstream investment, the potential impact of a successful breach grows exponentially. By ensuring that security teams have the same technological advantages as their opponents, the industry can better protect user funds, maintain trust, and foster sustainable growth. In conclusion, the coalition’s appeal represents a call for fairness and practicality in the deployment of AI safety measures. It urges AI laboratories to recognize the dual‑use nature of their technologies and to create pathways that empower legitimate security research while still safeguarding against abuse.
The hope is that through collaborative policy‑making, transparent access programs, and ongoing dialogue, the gap between defensive capabilities and malicious exploitation can be narrowed, ultimately strengthening the resilience of the cryptocurrency ecosystem and the broader digital economy.