CRYPTO FIRMS PUSH AI LABS TO LOOSEN SAFETY GUARDRAILS

The coalition of crypto and bitcoin companies contends that current AI safety guardrails create an asymmetry favoring attackers. Open-source security researchers face constraints on using advanced AI models for vulnerability discovery and threat analysis, while threat actors operate without similar restrictions. The companies have formally asked leading AI labs to grant security researchers early access to frontier models under controlled conditions. This would enable legitimate teams to identify and patch vulnerabilities before attackers exploit them. The request highlights ongoing tension between AI safety measures and practical security needs. While AI labs implement guardrails to prevent misuse, security professionals argue these protections can inadvertently disadvantage defensive cybersecurity work. Coinbase and Block, both major players in crypto infrastructure, joined the effort alongside three dozen other firms. The initiative suggests growing consensus within the industry that current safeguards require recalibration to balance safety concerns with legitimate security research requirements.