Every time a murder case or major crime happens, we often see comments like, “This is what happens when the death penalty is abolished.” But this is where clarity matters. What was abolished was not the death penalty itself, but the mandatory death penalty. That distinction is not small. It changes the entire meaning of the discussion. The death penalty still exists for certain offences, but judges now have more room to evaluate the full context of a case before deciding the appropriate sentence.
This is not about defending criminals. It is about ensuring that justice is not reduced to a fixed command without considering facts, intent, circumstances, evidence, and proportionality. In law, one word like “mandatory” can change the power of a judge. The same principle applies directly to AI. If our understanding of a matter is unclear, incomplete, or emotionally driven, the output from an AI agen will also reflect that weakness.
AI agents do not magically create clarity from confusion. They operate based on the instructions, context, parameters, and guardrails we provide. If we define the problem wrongly, the agen may execute the wrong solution perfectly. That is actually more dangerous, because a wrong decision made manually is already risky, but a wrong decision automated by an AI agen can scale very fast.
This is why human understanding must come before automation. Before we ask AI to decide, recommend, classify, approve, reject, escalate, or execute, we must first be clear about what we actually mean. Kejelasan is not just good communication. Kejelasan is governance. Kejelasan is risk control. Kejelasan is the difference between an AI agen that helps the business and an AI agen that creates a baharu problem at machine speed.
Whether in public policy, law, or artificial intelligence, the lesson is the same. Do not automate what you do not understand. Do not let emotion replace definition. And do not expect an AI agen to make the right decision when the human instruction behind it is already unclear.
