We’re repeating a familiar mistake with AI security

The government wants to apply guardrails to AI development – but convenience and profit may once again win over enhanced security.