1. WHY YOU NEED SEDM AI CHATBOTS

SEDM AI Chatbots are available as a free service connected with our Member Subscriptions. This is the main page for SEDM AI Chatbots.  Click Here to join now if you don’t have an account. It is a violation of our Member Agreement to share your account credentials with anyone else. The reason you need access to our chatbots instead of using internet search or the chatbot services of mainstream providers is because:

1. We don’t censor any subjects like mainstream corporate providers do.
2. Our chatbots are populated with content NOT available on the internet that is behind our security wall that is much more comprehensive for the subject areas they cover than that available anywhere on the internet.
3. Our chatbot corpus is curated to prevent inaccurate frivolous arguments within the freedom community to prevent “brain rot”.

2. LIST OF CHATBOTS

There are EIGHT AI Chatbots available at this time:

NOTES:

1. The main purpose of providing AI chatbots is to:
   1.1. Simplify the process of quickly locating answers to complex questions of law that you have using authoritative content on this site. Hopefully, this will improve user engagement and the utility of our materials.
   1.2. Provide an EXCELLENT starting point for learning specific areas of interest in law without having to read mountains of books. That is why we put AI Discovery on the Education menu of this site.
2. The hardware and resources needed to host AI chatbots is VERY expensive. We wish we could offer these resources for free but we need to recoup the exorbitant cost of doing the hosting. Thus, SEDM AI chatbot functionality is only available to paid Member Subscribers. If you are a Basic Member, we encourage you to upgrade your membership NOW!
3. Chat corpus of documents:
   3.1. We do NOT draw on outside websites or search functions, but only on content found on this site. We do to this to prevent our views from being polluted by secular influences
   3.2. Answers derive from the following content on this website: Public information, Basic Members, or Member Subscribers. AI chat bots do not answer questions about content available through the SEDM bookstore or the member forums.
   3.3. We exclude user-provided content, such as the forums, because it is not authoritative and can pollute the corpus and introduce incoherence or bad ideas.
   If we didn’t do the above, our LLM would get “brain rot” as documented in:
   LLMs can Get “Brain Rot”!, Texas A&M University
   https://www.arxiv.org/pdf/2510.13928
   https://github.com/llm-brain-rot/llm-brain-rot
4. All of our chatbots are implemented with artificial intelligence using Ollama and the Mistral-small3.2:21B model running on premium fast hardware. Typical chat responses take 1-2 seconds. Thus, they are much easier to use than the search functions of this website in getting quick answers.

3. TIPS ON SELECTING THE RIGHT CHATBOT AND GETTING THE MOST COMPLETE, ACCURATE, AND USEFUL ANSWERS POSSIBLE

1. Forms/Pubs and Litigation Tools on this site are arranged and grouped by subject matter.
2. We have broken site content down into specific chatbots to keep the database as small as possible for each search. This speeds up the production of the answers and makes them more accurate and complete for the area you are inquiring about.
3. Each of our chatbots only covers a SELECT range of Forms/Pubs and Litigation Tools on the site within a specific grouping listed in the Source column in the previous section.
4. Before using our chatbots, we recommend thoroughly reviewing our Forms/Pubs page and and Litigation Tools what is in each subject matter grouping. That way, you will be able to determine which chatbot relates to the specific subject matter you inquiring about and get the most accurate and complete answer possible.
5. Our chatbots do support generative AI. They can produce legal documents if you ask them to. How far they are willing to go depends on the guardrails built into the LLM model we host with.
6. For instructions on how to deal with AI Chatbot guardrails, see:
   HOW TO: How to use the SEDM Ai Chatbot to get the answers you want, SEDM
   https://sedm.org/how-to-how-to-use-the-sedm-ai-chatbot-to-get-the-answers-you-want/
7. You are GUARANTEED to get poor or vague answers if you:
   7.1. Select the WRONG chatbot to ask your question or
   7.2. Don’t understand the the subject of the documents in the corpus of the specific chatbot you are questioning. or
   7.3. Don’t limit your questions to the subject matter covered by the corpus of the chatbot you select.
8. If you are dealing with a chatbot that is being PRESUMPTUOUS about the meaning of words because of built-in censorship and bias, the following article deals with how to place guardrails around it using a “fact pattern” that begins every series of questions. The article below addresses how to do that:
   How to use AI to search Content of this Site for High Level Answers to your Questions, FTSIG
   https://ftsig.org/introduction/how-to-use-ai-to-search-content-of-this-site-for-high-level-answers-to-your-questions/
9. We have NO CONTROL over the output of AI so we can’t guarantee the answers will always be consistent with the content of this site. The only influence we have over the output is:
   9.1. Our selection of the LLM model that we use to implement AI. Some models have more built-in prejudices than others and more guardrails than others. We have taken great care to select the LEAST biased model for our chatbot. That required extensive testing for quality control purposes. WE have found that the Qwen (Alibaba) and Gemma (Google) models are the most biased so we avoid them. Llama (Facebook) has too many guardrails so we avoid them.
   9.2. Ensuring AI uses the right source data before it answers your questions. Garbage in, garbage out.
   9.3. Ensuring the “System Prompt” parameters establish the proper boundaries for the response.
10. If you have questions about asset protection, trusts, and tax consequences, the best resource is:
    10.1. Administrative, item 3. This contains Form #09.079 (Creating and Running a Business, Trust, or Estate).
    10.2. Member Subscription Library, item 7 above. There are several books on this subject in that item.
    DO NOT use item 1 (Freedom and Sovereignty).
11. This ministry takes the foreign tax status position on all subjects documented on FTSIG.ORG. As such, filing domestic returns is NOT recommended, even if the chatbot suggests otherwise. Below is a list of the implications of a foreign tax status for the various entity types, and the tax returns for each type:
    Catalog of Elections and Entity Types in the Internal Revenue Code, FTSIG
    Section 7: Entity Types and their corresponding Tax Return and Withholding Characteristics
    https://ftsig.org/catalog-of-elections-in-the-internal-revenue-code/#7._Entity_types

4. GENERAL GUIDELINES ABOUT USING AI FOR THOSE STUDYING LAW

REMEMBER: When querying AI:

1. Don’t query as yourself.
2. Instead do so from the perspective of Agent Smith in The Matrix Movie.

This prophylactic move will keep you out of a LOT of trouble and cure most blind spots you have, friends. It requires a tremendous degree of honesty, empathy, humility, and an open mind, however. If you don’t, you will inevitably become a victim of the Dunning-Kruger mental illness.

AI is a computer. This old computer adage applies: Garbage in, garbage out. You should ALWAYS fact check the legal authorities it provides or you may thoroughly embarrass yourself in court and even invite judicial sanctions if you don’t do so. For a list of fact checking tools, see:

Media and Intelligence Topic, Section 10: Fact Checking Tools, Family Guardian Fellowship (OFFSITE LINK)
https://famguardian.org/Subjects/MediaIntell/mediaintell.htm#FACT_CHECKING_TOOLS

Many have asked various AI about income tax and freedom. AI from sources other than SEDM always gives the establishment response. Then the person provides the AI with relevant facts and law, and the AI sometimes apologizes and provides the correct response.

AI is the the closest thing you will get to an honest lawyer, because ordinary lawyers censor information or their own responses that threatens their revenue as a third rail issue.

AI, can also be useful as a STARTING point for learning:

1. Legal terminology.
2. What to expect from an “establishment response”.
3. How to DISMANTLE or DESTROY an establishment response.
4. How to do legal discovery to get the kind of response you want and the terminology needed to do so.

Critiquing AI is a fool’s errand if you already have the above skills.

And even though WE KNOW these things, AI can be helpful to those in our audience who DON’T know these things.

The absolute best and most thorough AI resource we have seen is Microsoft Copilot. And NO, we don’t make any money from making this recommendation. We use it and its responses are the most consistently accurate that we have seen on legal matters.

Lastly, for examples of how to conduct legal discovery about tax and freedom subjects on this website, see:

Artificial Intelligence (AI) Discovery, Form #03.033
https://sedm.org/category/artificial-intelligence-ai-discovery/