FAQ

Curious about how Deep Discourse works? Here are answers to common questions about our AI-powered philosophers, their knowledge sources, and how to get the most from your conversations.

How to Chat with AI Philosophers

1. Select an AI philosopher on the home page
2. A text-messaging interface will appear
3. Type you first question and press the send button to start the chat

General Questions

• What is Deep Discourse?
  Deep Discourse is a website that hosts AI philosopher characters you can chat with.
• How does Deep Discourse work?
  We align specialist AI models on the writings of each philosopher and provide a chat interface so you can interact with them
• Which philosophers can I chat with?
  The full list is accessible from our home page. New names are added weekly. You can always request a name you don’t see.
• How accurate are the AI-generated responses?
  We work hard to ensure the responses represent the original written words, paraphrased. The responses will usually include a citation so you can verify the responses.
• Can I use this for academic research?
  You are welcome to use it for research, it can be especially useful for finding citations on a topic before deeper study of the original work.
• How does this compare to reading the original texts?
  Our goal is to discover new ideas that align with, but are not directly addressed, in the source text. We enjoy reading as much as anyone and are not trying to replace that.

AI & Technical Questions

• How is the AI trained to mimic each philosopher?
  We use up-to-date LLM models that are domain-aligned with source text voice, then use RAG to generate responses that are true to, and even cite, the source material.
• What sources do the AI philosophers reference?
  Each AI philosopher is trained on a corpus of their primary works. The exact texts are noted on each AI philosopher’s page.
• Does the AI provide real historical responses, or are they generated?
  The responses are generated conversationally, balanced towards accuracy over creativity.
• Can the AI learn from user interactions?
  At this time, the AI only has context from the start of your conversation. In the future we may use recorded human interactions to improve the model behavior.

Philosophy & Learning

• Which philosopher should I start with if I’m new to philosophy?
  A chat with Plato is a great start! As A.N. Whitehead said: “The safest general characterization of the European philosophical tradition is that it consists of a series of footnotes to Plato.”
• Can I compare answers from different philosophers?
  Not yet, but we’re working on that!
• Does the AI cover religious and ethical topics?
  Yes, the AI generates responses that are faithful to their source texts which cover all the most difficult discussions of human life.

User Experience & Features

• Is Deep Discourse free to use?
  Yes, we want to ensure you can chat with any AI philosopher for free.
• Will you add more philosophers in the future?
  We are frequently adding new names as we try to cover a wide swathe of human history and geography.
• Can I request a specific philosopher to be added?
  Definitely – please visit the requests page.
• Is there a mobile app version of Deep Discourse?
  This website is mobile-optimized as much as possible. In the future we might create a smartphone app but for now the website allows us to build faster.

Monetization & Community

• Are there any premium features?
  No, all features are free, but as an Amazon Associate we earn commissions from qualifying purchases made through links we provide.
• How can I support Deep Discourse?
  Please provide feedback on your interactions with the AI philosophers and request new ones!
• Can I contribute by improving the AI or submitting content?
  We can use help with sourcing texts, verification, and model training. Please reach out if you would like to help.

Privacy & Ethical Concerns

• How do you prevent misinformation in AI responses?
  Models are given some limited novelty (temperature) or else they sound like robots. But they are also instructed to instructed to cite source texts so you can fact-check the models for yourself.
• Is my data safe when chatting with the AI?
  It is truly anonymous unless you tell us your name. But you should assume it’s recorded and used by us and our service providers to improve the services.
• Does the AI have biases? How do you handle them?
  Our objective is to make the AI philosophers extremely biased towards their own source materials. The problematic bias is that our LLMs start with the latest fad of western ethics embedded and we have to work hard to bend them into the mindset of the philosopher they represent, no matter how offensive their worldviews might seem today.
• Can the AI be used for debates and discussions?
  We’re actively working on that! currently the chats are in a question/answer format that doesn’t try to argue or maintain long contexts.