Where Is the Panic Button?

Why This Post Exists

I was using Codex recently when it started heading in a direction I did not like. Nothing dramatic happened. No glowing red screen. No sci-fi disaster. Just a coding agent confidently marching to its own beat, in the wrong direction while I sat there thinking: Wait. That is not what I meant.

Then came the practical question: How do I stop this?

There was a Stop button. I pressed it. The run stopped. I breathed again. And then I asked myself: Why did I start something before I knew how to stop it?

That is not a criticism of Codex. It is not fear mongering. I knew better. It is basic safety thinking: Before you start the machine, know where the stop control is.

Stop Is Not Panic

In ordinary chat, the Stop button is usually enough. If the answer goes sideways, stop it, redirect, or start over.

But agentic tools are different.

A coding agent may edit files. A workflow agent may call tools. A publishing agent may create content. An email agent may prepare messages. A data agent may query or transform records.

Once a tool can act, operating hygiene demands that the human supervising it needs to know:

1. How do I stop it?
2. What did it already do?
3. How do I undo or recover?

The Thing That Surprised Me

After this happened, I asked a few people who use GenAI: Where is the panic button?

Most, like myself, had not yet thought about it. Why? We have not been taught that stopping the tool is part of using the tool.

In safety culture, there is a familiar idea: Stop Work Authority. If work appears unsafe, people are allowed to stop it. In some cases, they are expected to stop it.

That idea belongs in GenAI, not because AI is dangerous in some cartoon-villain way, but because the pattern is familiar: If the work is going wrong, you are allowed to stop the work.

Maybe more than allowed. Sometimes, we are obligated.

The Industry Is Moving in the Right Direction

The good news is that the industry is already working toward better controls. We are seeing more attention on:

• Undo
• Rollback
• Approval gates
• Read-only modes
• Sandboxing
• Observability
• Human override
• Action logs
• Tool boundaries

I would also love to see more tools support single-step mode. One proposed action. One approval. One execution. One result. Then the next step.

Until Then, We Still Have Stop

A Stop button is not enough if users do not know it is there. It is not enough if they do not know when to use it. It is not enough if they do not feel authorized to press it.

A Stop button should not feel like panic. It should feel like a brake pedal, an emergency stop, or a circuit breaker. It is there so we can safely do more interesting things.

A Simple GenAI Safety Briefing

Every GenAI rollout should include a small safety briefing:

• Know how to stop the tool.
• Know what the tool can change.
• Know when it will ask before acting.
• Know where to review logs, diffs, or outputs.
• Know how to undo or roll back changes.
• Know who to tell if something goes wrong.

That is not bureaucracy. That is trust infrastructure.

Before you hand work to an agent, know how to stop the agent.