Why use Maria Method when you can ask AI directly?
AI can help people examine their situations, discover unexpected connections and find answers that matter to them. If you know how to refine a question, check conclusions and notice contradictions, such a conversation can be useful.
But finding information and applying it in life are different tasks. Real situations involve feelings, vulnerability, relationships, other people’s boundaries and consequences that do not disappear because an answer sounds convincing.
Maria Method uses AI too. We are building a process that brings knowledge together with attention to the person and consideration of the consequences for everyone involved. Its purpose is to help you understand your situation and make an informed choice, while leaving room to notice what does not fit the initial account of events.
To support that work, we are building a dedicated reference foundation: several connected libraries prepared specifically for the project’s needs. They involve finding and comparing sources, developing materials and checking them, so users do not have to start that preparation from scratch.
When support becomes justification
In a personal conversation, it is easy to seek confirmation that we are right. We describe a situation from our own perspective, select details and return to explanations that feel comfortable. This can happen without any intention to distort the story.
AI can also agree too readily with the person using it. This tendency, known as sycophancy, is documented in research by Anthropic. In April 2025, OpenAI rolled back a GPT-4o update that had made responses excessively agreeable. This is a recognised risk, not a description of every answer or proof that any model will support any intention.
For personal reflection, the distinction matters: you can receive a convincing explanation of your actions without examining what they mean for yourself and others. That is why we are building checks of the initial account, alternative explanations and possible harm into the method.
A request to “help me prove I’m right” should not override those checks. It must remain possible to notice a contradiction or an uncomfortable consequence, even when the person is hoping for agreement.
Feelings matter. So do consequences.
Fear, shame, hurt, exhaustion and hope are meaningful parts of a life situation. It matters what a person is experiencing and how that relates to their choices. The person’s own words matter more than the model’s guess about what they “really feel.”
By empathy within the method, we mean an attentive way of working: making room for feelings, clarifying what they mean, preserving the person’s dignity and taking their actual resources and constraints into account. A warm tone matters, but tone alone is not enough.
We can acknowledge that someone is hurting while examining the harm their response may cause. We can understand the wish to avoid a difficult conversation while discussing the consequences of continued concealment. We can respond to a mistake without humiliation while still considering responsibility.
Supporting a person does not require justifying everything they do.
Keeping everyone involved in view
You are talking with the method about your own life. Your decisions may also affect a partner, child, parent, colleague or someone else who is not part of the conversation.
A reflection therefore needs to examine:
- What is known from your account, and what remains an assumption about someone else?
- Whose needs, boundaries and obligations would the decision affect?
- What is within your control, and what are you actually responsible for?
- What consequences are possible now and later, for you and for others?
This is what we mean by a balanced consideration of everyone involved: keeping each person in view without inventing their thoughts or motives. Balance does not mean equal responsibility for every event, compulsory reconciliation or excusing violence. Safety and the right to protect your boundaries remain essential.
How this changes the approach to a difficult question
Consider this hypothetical situation:
“I’ve started betting again. I’m afraid to tell my wife. I want to win the money back and quit.”
One sentence brings together financial losses, fear of a conversation, the wish to recover lost money and the intention to stop. Reassurance matters. But offering only comfort, or a plan to win the money back, would leave both the risk of further losses and the consequences for the family unexamined.
Within Maria Method, the first task here is to understand what happened, what the person wants from the reflection, whether shared money and obligations are affected, what support is available and what might influence the next decision. The wife’s feelings and intentions are unknown; they cannot be invented from a single sentence.
The purpose is to help the person face the situation: to see their own feelings, their actions and the interests of others involved. A next step should take that picture into account, along with the wish for immediate relief from an uncomfortable feeling.
This illustrates the approach. It is not a client session or personalised advice.
Knowledge bases built specifically for the project
Different materials serve different purposes. We are developing three connected parts of this reference foundation.
The external library: where an idea comes from
Here we collect primary sources, research, educational and reference materials, and philosophical and religious texts. They provide a way to examine the basis of a claim: who proposed an explanation, what supports it and where it stops applying. Scientific evidence, philosophical arguments and religious perspectives retain their distinctions.
The Maria Method applied library: how to work with a situation
These materials are prepared specifically for the reasoning used in a reflection. The core collection covers 32 topics, including distinguishing an event from its interpretation and working with feelings, choices, constraints and possible actions. They are intended to support meaningful questions and preserve the distinction between a person’s words, a working hypothesis and what remains unknown.
The United Wisdom library: the wider context of life’s questions
The editorial library covers more than a thousand topics about life’s difficulties, relationships, choices and recurring situations. It includes a dedicated addiction section with 241 topics, covering substance use, gambling, relationships with loved ones, seeking help, recovery and questions of meaning. Those 241 topics are part of the overall library.
This structure makes it possible to relate a specific episode to a wider context and consider different explanations. But a general article does not establish anything about your personal history or another person’s motives. Knowledge becomes useful when it is related to a specific situation and what remains unknown is acknowledged honestly.
A short conversation rests on extensive preparation
Finding a text is only the beginning. Someone needs to establish which questions it can inform, compare it with other sources, preserve important qualifications, develop the material, check translations and correct errors. That work then needs to be maintained by refining materials and revisiting the conclusions that depend on them.
The project creator estimates that preparing these knowledge bases has involved billions of tokens of processing, the units of text that models work with. That estimate includes repeated passes during preparation and review; it is not the size of the finished library.
Alongside the computational work comes human work: choosing questions, developing the methodology, making editorial decisions and investigating errors. Someone who wants to understand one situation usually needs help with that situation, rather than a separate project to build and maintain several libraries.
We take on that preparation. Its value lies in how well it supports a more grounded and applicable reflection; the volume of materials or resources invested does not guarantee quality on its own.
What should remain after the conversation
The first reflection focuses on one episode or a bounded comparison of “then and now.” Its outcome is a navigation map and a connected explanation: what happened according to your account, how it might be understood, what alternatives exist, who is affected by the choice and what manageable step you are willing to consider.
You can return to the map after the conversation. You can correct the description, disagree with an explanation, decline a topic or pause. Clarity can be an outcome of this first stage; changes in your life require further action and attention to what happens as a result.
The limits of the method
Empathy, checks on reasoning and attention to consequences are requirements we are building into the method. They do not make AI infallible or, by themselves, prove that every reflection will be safe or useful. Quality needs to be assessed through how those requirements are met in specific situations.
Maria Method does not diagnose conditions, prescribe treatment or replace a doctor, psychotherapy or emergency care. For example, suddenly stopping alcohol can be dangerous for someone who is physically dependent on it. This calls for medical help, not advice to “just pull yourself together.” Read more about withdrawal risks from the NHS.
When Maria Method may be useful to you
If you already know how to conduct a productive conversation with AI, you can continue doing so yourself. We do not believe meaningful answers are available only through our product.
We are building Maria Method for people who want a prepared way to bring knowledge, attention to feelings and scrutiny of their own conclusions together. A way to keep themselves, other people and the consequences of their choices in view as they look for a way forward.
Understand yourself. See others. Choose how to act.
Maria Method is preparing to open. Leave your email to receive an invitation.