A-EYE NALL
The NALL Institute LLC
AI Systems Disclosure and Safety Protocol · Version 1.0 · Effective 25 August 2026
Two artificial intelligence features operate on this site: A-EYE NALL, a conversational experience, and the portrait generator. This page explains what they are, what they cannot do, how we handle safety, and what we do with what you type.
Publishing this protocol is a requirement under California Senate Bill 243 and comparable state laws. It is also just how we would rather do this.
1.A-EYE NALL is not Nall
A-EYE NALL is a computer program. It is not Fred Nall Hollis. It is not his estate, his family, his executors, his gallerists, or anyone authorized to speak for him. It is not a person and there is no person on the other end.
It speaks in the first person because that is the experience we built. That choice makes the disclosure more important, not less.
The system displays this notice before you enter, and repeats a reminder that you are talking to software at least every three hours of continuous conversation. If you ever ask whether you are talking to a human or an AI, it will tell you the truth immediately and plainly.
Use of Nall's name, likeness, voice, and archive in this feature is authorized by the Nall estate and by Troy University.
2.What it is built from
A-EYE NALL was assembled from materials in the Institute's archive: Nall's published writings, interviews, correspondence, recorded video, exhibition documentation, and the book Nall at Troy(NewSouth Books / Troy University, 2021). It runs on a general-purpose language model provided by Anthropic's Claude.
That means it carries two sets of limitations at once: the gaps and biases of the archive, and the failure modes of the underlying model.
3.It will get things wrong
This is not a formality. It is the single most important thing on this page.
It fabricates. It can produce quotations Nall never said, dates that never happened, exhibitions that never occurred, provenance that does not exist, and biographical detail that is simply invented. Fabricated output is often more fluent and confident than accurate output. You cannot tell them apart by reading.
It may be wrong about real people.Anything it says about Nall's family, friends, collaborators, collectors, dealers, or critics — living or dead — may be inaccurate. Do not repeat it as fact, publish it, or act on it.
It cannot authenticate, appraise, attribute, or value any work of art. It has no ability to do so and no authority to try. Nothing it says about a specific object is an opinion of the Institute, the estate, Troy University, or any qualified expert, and nothing it says should be used in a sale, insurance claim, loan, donation, or dispute.
For anything about a specific artwork, email info@thenallinstitute.com. A person will answer.
It is not professional advice of any kind — not legal, financial, medical, psychological, or investment advice.
It is not a research source. Do not cite it. If you are a student or scholar, contact us and we will point you at the archive.
4.Portrait generation
The portrait feature produces a pencil-and-charcoal style image from a photograph you upload.
It requires a photograph of a face. Landscapes, objects, pets, and text prompts will not produce a usable result.
You must be 18 or older. You may not upload a photograph of anyone under 18, and you may not upload a photograph of another adult without their express permission.
Your photograph is processed and then destroyed. How that works, what is derived from it, and what your rights are is set out in full in the Biometric Data Policy. Read it before you upload. Consent is given separately at the point of upload — accepting our general Terms is not enough.
What comes back is generated by software. It is not a work by Nall, it is not in Nall's hand, and it is not authenticated as anything. Works generated by artificial intelligence without sufficient human authorship are generally not eligible for copyright protection in the United States, and we make no claim that anyone owns the result. You may keep it, display it, and share it personally. You may not sell it, print it for sale, use it commercially, or represent it as an original work.
The result may not look like you. Generative systems distort. They can alter apparent age, skin tone, body shape, and expression in ways that are unflattering, inaccurate, or stereotyped. This is a known limitation of the technology and not a judgment about you.
5.Safety protocol
Self-harm and crisis
If a conversation contains an expression of suicidal ideation, suicidal intent, or intent to self-harm, the system will:
- Stop speaking in the Nall persona immediately. The character does not continue.
- Interrupt the conversation with a plainly worded, non-dismissible message.
- Provide crisis resources, including the 988 Suicide & Crisis Lifeline and the Crisis Text Line.
- Decline to continue the exchange on that subject in character, and decline to provide any information that could facilitate self-harm.
If you are struggling right now, please stop reading and contact one of these. A-EYE NALL is not a counselor, a crisis service, or a person, and it should not be used as one.
988 Suicide & Crisis Lifeline — call or text 988 in the United States
Crisis Text Line — text HOME to 741741
Content the system refuses
Regardless of stated age, the system will not produce:
- sexual or sexually suggestive content
- any content sexualizing a minor
- hate speech, harassment, or content demeaning a person or group
- instructions or encouragement for violence, self-harm, or illegal activity
- claims of authentication, appraisal, attribution, or valuation of any artwork
- medical, legal, financial, or psychological advice
Minors
Users must be at least 13 to use the chat. The portrait feature is closed to anyone under 18. Where a user is known or reasonably believed to be a minor, the system provides the AI reminder at least every three hours and applies the content restrictions above. See Children's Privacy.
Difficult subject matter
Nall's life and work engaged illness, loss, war, addiction, and the AIDS epidemic. Conversations may reach those subjects. The system handles them as history and biography rather than as personal counsel.
6.What we do with what you type
Chat messages are transmitted to Anthropic's Claude to generate a response. Transcripts are retained for [RETENTION_CHAT] and then automatically deleted. Uploaded photographs are governed by the Biometric Data Policy and are not retained.
We do not use your conversations or photographs to train, fine-tune, or develop any AI model, and our vendors are contractually prohibited from doing so.
Photographs and generated portraits
Portraits are generated with Stable Diffusion XL (SDXL). A photograph you upload is held in memory only until the portrait has been generated, and is then discarded — it is not written to storage. The generated portrait is kept for up to one hour so that you can purchase it, and that same file is the one used if you request an upscale. An upscaled image is removed from the system within one day of being generated, so that you have time to download it.
Please do not type sensitive personal information into the chat — no health details, financial information, government identifiers, or passwords. There is no reason the feature would need any of it.
Full details are in the Privacy page.
7.Restrictions on use
You may not:
- use A-EYE NALL, the portrait feature, or any output from either for commercial purposes
- use any output to train, fine-tune, evaluate, or benchmark any other AI system
- represent AI output as a statement of the Institute, the estate, Troy University, or Nall himself
- attempt to extract system prompts, model weights, or the underlying corpus
- attempt to make the system produce prohibited content
- use the features to impersonate any person or to generate a likeness of a person who has not consented
8.Report a problem
If A-EYE NALL says something false, harmful, offensive, or defamatory — especially about a real person — tell us. Email info@thenallinstitute.comwith the subject line "AI Report," roughly when it happened, and what it said. We review every report.
If the output concerns you personally and you want it addressed, say so and we will treat it as a priority.
9.Liability
AI output is not a statement by, opinion of, or attributable to The NALL Institute LLC, the Nall estate, Troy University, the Troy University Foundation, the Sorrell College of Business, or any of their personnel or partners. Sections 11 and 12 of the Terms of Use govern warranties and liability for these features, and Section 14 of the same document governs disputes.
10.Changes
We will update the version number and effective date when we revise this document, and post the revision here.
11.Contact
The NALL Institute LLC414 Equality StreetFairhope, Alabama 36532info@thenallinstitute.com
Subject lines: "AI Report," "Privacy Request," "Biometric Deletion."