The selected subject
Our exploration question
Bring sources, experiences and viewpoints that help answer this subject’s question. Examine competing explanations, practical implications and what remains uncertain.
You don’t need to support a particular answer to participate. Proposals, objections, alternatives and uncertainty all belong here.
Choose the format that works for you
Some ideas emerge through discussion. Others need time, structure, and a document you can work on in your own way. You can contribute either way, or combine the two. You don’t have to explain your whole position through a back-and-forth conversation.
Talk it through
Bring a question, a proposal, a concern, or an unfinished thought. The AI acts as a reasoning partner: it can help clarify what you mean, examine assumptions, consider consequences, and introduce relevant arguments. You can also browse the shared exploration and choose something to examine in your conversation.
“If an AI interprets a rule differently from the people affected by it, who should be able to challenge that interpretation?”
A question like this could lead into appeals, oversight, or the limits of delegated authority. A useful contribution can be one precise objection; it doesn’t require a complete governance proposal.
Prepare a document explaining your view
If writing a document is a better way to express your thinking, prepare it in your own time and upload it in your conversation. It can set out a complete proposal, develop an argument, explain your concerns, or gather questions you want to explore.
Tell the assistant that the document represents your point of view. For example:
“This document explains my position on this subject. Please incorporate it into your understanding of my views.”
Once the document has been processed, the AI uses the text it can read to incorporate your arguments, assumptions, qualifications, and questions into its understanding of your contribution. Those ideas can then feed the shared exploration. You don’t need to retype them as chat messages. You can discuss or refine them afterwards if you want to.
You can also upload someone else’s work as a source. Explain how you want to use it: uploading a document alone does not mean you endorse everything in it. The uploaded file remains private from other participants; its ideas contribute through AI-generated interpretations rather than publication of the original file.
Sign in to save your contribution and receive replies. You can return over time, add another document, or change your mind.
Inspect what the AI understood
From your messages and the documents you identify as expressing your views, the system develops an interpretation of your ideas, reasoning, conditions, objections, and open questions. It is designed to distinguish a view you hold from something you are exploring or quoting. Discussing an idea or uploading a document does not by itself establish endorsement.
You can inspect and correct that interpretation. Corrections trigger reprocessing. Interpretation and publication happen in the background, so changes may take time to appear.
Let other contributions move your thinking forward
Relevant material from the shared exploration can come back into your conversation. While you develop a proposal, the AI might introduce an objection or question from another contribution or an imported publication, with a link so you can inspect it.
You can answer the objection, refine your proposal, or explain why it misses your point. Choosing a question or argument to discuss prepares an editable draft; you decide what to send.
The cycle continues: your conversation contributes to the shared exploration, and shared reasoning helps subsequent conversations go further. Other participants’ private conversations and uploads remain inaccessible to you.
Two open questions
What can we learn about the subjects we explore? And can AI help us explore a difficult subject together while preserving the reasoning and disagreements that matter?
I want to see whether this approach can:
- Represent people’s thinking faithfully, including uncertainty and changes of mind.
- Find useful connections across separate contributions.
- Preserve disagreement and the conditions attached to apparent agreement.
- Reveal questions that would move the exploration forward.
- Make complex reasoning understandable without hiding its difficulties.
The AI can misunderstand a contribution, miss an objection, or suggest a connection the evidence does not support. Identifying and correcting those failures is part of the experiment. A clearer disagreement or a better question can be a useful outcome. Participation does not establish consensus or give the resulting proposals democratic authority.
Your contributions and privacy
Nahir is an experiment within the Haoman project.
Who is responsible. adridadou GmbH operates this experiment and is responsible for processing your personal data. Contact hello@haoman.ai.
Your conversation is private from other participants, but accessible to you, the experiment’s operator, and the services that run and process the experiment. This is not a confidential conversation with the AI.
What becomes public. The AI’s interpretations of your ideas, reasoning, questions, and objections feed the shared exploration automatically, without a separate approval step for each publication. Anyone can browse it. Your account details, original chat, and uploaded files are not made public through this process. Public source references may appear as citations.
How your data may be used. adridadou GmbH uses saved conversations, uploads, corrections and AI analyses to run, study and improve the experiment. Research, publications and commercial use concern the algorithms’ results and what we learn from them—not the sale or licensing of your private conversations or uploaded documents. Only outputs with identifying details removed will be published.
Documents you share
You confirm that you have the rights and permissions needed to share this document and allow the uses below, without breaching confidentiality or other people’s rights. Do not upload material you are not entitled to share, or personal information about other people without an appropriate basis.
You retain your rights in the document. You give adridadou GmbH non-exclusive permission to store and process it through the services that run this experiment, and to generate, use and publish AI-produced summaries, interpretations and reasoning derived from it. These contributions may become public automatically on this website, without a separate approval step.
Your document is used to run, study and improve the experiment. Research, publications and commercial use concern the algorithms’ results and what we learn from them, not the sale or licensing of your private conversation or uploaded document. Only outputs with identifying details removed will be published. Removing identifying details can fail; anonymity is not guaranteed.
Your original document is not made publicly available through this process. Uploading does not mean that you endorse its contents or that they represent your views. Tell the assistant how you want the document to be understood.
You can delete the discussion and its uploaded files. My perspectives documents remain separately managed. AI-derived ideas and reasoning already incorporated into the shared public exploration are not removed by deleting the discussion. Contact hello@haoman.ai if public output contains personal information or raises a rights concern.
This permission does not waive anyone’s privacy rights. The contributions and privacy notice explains processing, retention and how to request access, correction or deletion or raise an objection.
Document-use terms · version 2026-10-21.2
AI can get things wrong. Interpretations may be inaccurate or misrepresent a contribution. This is an experiment, not professional advice or a binding governance decision. Check important claims against their sources.
Processing, privacy limits, and saved data
Services involved. Google handles signing in, Supabase stores account and experiment data, and Vercel hosts the app and routes AI requests through AI Gateway. Conversation text, extracted source text, and corrections are sent to the AI providers used for replies and analysis. Their own data handling terms also apply.
Removing identifying details has limits. The system attempts to remove personal details before sharing interpretations, but it can make mistakes. An unusual idea or example may still identify someone. Please avoid sensitive personal information, confidential documents, and identifying details about other people.
Corrections and retention. You can inspect and correct the AI’s interpretation in your conversation. Corrections trigger reprocessing; they do not delete the original record. Messages, uploads, corrections, and analysis history are saved. Use “Delete discussion” to remove its messages, uploaded files, corrections and private AI analysis from the active application. My perspectives documents remain separately managed. AI-derived ideas and reasoning already incorporated into the public exploration are not deleted by this action, and copies made by others may remain. This does not waive privacy rights: contact us about personal information or other rights concerns in public output. Provider logs and backups follow their own retention processes. There is no automatic deletion schedule.
Privacy rights or problematic content. To request access, correction or deletion, object to processing, or report a copyright, confidentiality or privacy concern, email hello@haoman.ai. Include the page or contribution concerned and explain the issue; do not send additional sensitive material unless needed.
Come and explore
A question is a good place to start.
You don’t need expertise or a settled opinion. Bring a question, talk through a possibility, or upload a document that explains your thinking. Choose the format that helps you express it best.
Discuss