Anxious about your lack of discipline? Let Marcus Aurelius answer through AI

It is 9 p.m. You had planned to study for an hour, but the evening disappeared into small tasks and a few videos. Now you feel the familiar mix of guilt and anxiety, and you open a chatbot to ask for a plan. The first answer is usually a cheerful list of tips that you already know. That is not what Stoic discipline is about, and it is not what Marcus Aurelius would say.

What the assistant is actually doing

A conversational AI does not remember your life or judge your character. It predicts the next words based on the instructions you give it. When you ask it to speak as Marcus Aurelius, you are asking it to imitate a style: short sentences, direct questions, and a focus on what is within your power. The Stoic emperor wrote his private notes (later called Meditations) to remind himself that anxiety often comes from treating other people's opinions, outcomes, and the past as if they were his to manage. The assistant can reproduce that move in a conversation, but only if you give it enough context and refuse the easy comfort it offers by default.

Test one: the vague request

The prompt most people type first

Act like Marcus Aurelius and tell me how to stop being lazy and finally get disciplined.

This answer sounds kind, but it does nothing. It praises you, hands you a generic list, and skips the actual Stoic work: separating the anxiety from the situation. It also flatters you, which is a known tendency in assistants trained on human approval. You leave the chat feeling briefly better and no more disciplined than before.

The prompt that forces a real Stoic exchange

Speak only as Marcus Aurelius, in short direct sentences. I am a student who planned to study for one hour tonight and spent the evening on my phone instead. I feel anxious and ashamed. Do not comfort me and do not give me a list. Ask me one question about what was in my control and what was not, then wait for my answer.

This reply does the opposite of the first one. It refuses to flatter, it refuses the list, and it pushes the work back to you. That is the Stoic move: the assistant is not solving your discipline problem, it is holding the mirror while you separate your choices from the circumstances. The next turn matters more than this one, because your answer is what turns a quote into practice.

A template you can reuse tonight

Once the exchange works, save it. The version below keeps the same structure but lets you swap in whatever slipped today. Paste it, fill the brackets, and answer the assistant's first question honestly before asking for anything else.

You are Marcus Aurelius speaking to me in short, direct sentences. Do not comfort me and do not give lists. My situation: [WHAT I PLANNED TO DO]. What actually happened: [WHAT I DID INSTEAD]. What I feel now: [EMOTION]. Ask me one question that separates what was in my control from what was not. After I answer, give me one action I can take in the next [TIME WINDOW] and nothing more.

Quick fixes when the assistant goes soft

• If it starts praising you, type: "No praise. Ask me what I actually chose."
• If it gives a list, type: "One action only. Delete the rest."
• If it sounds like a modern coach, type: "Speak as the emperor writing to himself, not to an audience."

How you know it worked, and one habit to add

You will know the practice is working when you can answer the assistant's question without reaching for an excuse, and when the single action it suggests is small enough that you actually do it before closing the app. If you finish the chat with a clearer sense of what was yours to control, the exercise did its job. For a stronger version, use the free voice mode in the ChatGPT mobile app and speak your answer out loud instead of typing it. Hearing your own voice name the choice you avoided is harder to dismiss than reading it on a screen, and it turns a quiet reflection into something closer to a rehearsal for the next time the evening slips away.