Case Study
ETHER 1227
Equitable Testimony through Human-Empowered Response
An assistive communication application built for the courtroom, where testimony is irreversible and authority cannot be ambiguous.
Case Study
Equitable Testimony through Human-Empowered Response
An assistive communication application built for the courtroom, where testimony is irreversible and authority cannot be ambiguous.
ETHER 1227 is an assistive communication application built for live court proceedings. It is predominantly rule-based: the large majority of what it does is deterministic matching against the litigant's own prepared testimony. It uses limited, constrained AI assistance only when rules cannot match a question, and Human Authority Architecture is its governance and reporting framework. The litigant reviews and approves every word before it is spoken.
ETHER 1227 belongs to the established category of augmentative and alternative communication (AAC) devices, long accepted in court. It extends that category to litigants whose barrier is neurological processing rather than physical speech impairment.
It is designed for people with documented disabilities that affect courtroom communication, including Autism Spectrum Disorder, Traumatic Brain Injury, aphasia, PTSD with dissociative features, and selective mutism. These are people who hold the full truth of their own case but cannot deliver it in the moment the court demands it, whether because of processing speed, difficulty with spontaneous speech, sensory overload, or an affect that is misread as evasiveness. ETHER 1227 exists so their testimony can be heard, accurately and in their own voice, without their giving up control of a single word.
A courtroom is the clearest test of human authority there is. Testimony is sworn, irreversible, and consequential. The stakes can be a person's parental rights, their liberty, or their livelihood. Any system that assists with testimony, including one that uses AI, must never originate testimony, never fabricate, and never speak on a person's behalf without their explicit, reviewable consent. ETHER 1227 was built to that standard from the first line.
ETHER 1227 follows a four-step cycle. It listens to the courtroom and identifies when a question is directed at the litigant. It matches that question against the litigant's own prepared testimony and documented evidence. It presents a proposed response on screen for the litigant to read and consider. And it speaks the response aloud in the litigant's own voice only after the litigant approves it. The litigant can approve, modify, or decline, and there is no time limit on that choice.
ETHER 1227 does not change how testimony is given. The litigant is sworn in exactly as in any proceeding. Because the litigant prepares, reviews, and approves every word, the responses delivered through the device are the litigant's own sworn testimony, spoken in their own voice. The device serves as a reasonable accommodation, in the same tradition as the AAC devices courts have long permitted, so that a litigant who would otherwise be unable to testify can give testimony that goes on the record. Admissibility remains the court's determination, and the complete Testimony Record Report is provided to support it.
ETHER 1227 is predominantly rule-based. Most responses are handled by deterministic rules, with no AI involved. Where AI is used, it is a narrow, constrained assistant, not the engine. Human Authority Architecture governs the entire cascade through its ACT model, in which authority only ever moves toward the human, never away.
Deterministic matching against the litigant's prepared testimony. No AI is involved. This tier handles the large majority of responses. It is Stability of Systems: predictable behavior that holds under pressure.
Used only when the rules cannot match a question. AI works against the litigant's case data alone, under strict constraints that prevent fabrication, speculation, and legal conclusions. It is Stewardship of Intelligence: capability held under tight custody.
The litigant reviews and approves every response before it is spoken. There is no time limit on review. It is Authority of Humans: final, absolute, and impossible to bypass.
ETHER 1227 uses artificial intelligence for limited assistance only, and only when its deterministic rules cannot match a question. The AI never originates testimony, never fabricates facts or evidence, never draws legal conclusions, and never speaks without the litigant's explicit approval.
All AI assistance operates inside the Human Authority Architecture governance framework, and the litigant approves every word. Humans Lead. Machines Assist.™ In ETHER 1227 that is not a tagline. It is the system, enforced in the architecture itself.
Reporting is part of Human Authority Architecture. ETHER 1227 generates a Testimony Record Report documenting every response delivered through the application: the full text of every question and approved response, the source evidence each response was drawn from including exhibit and document references, the processing method used, the litigant's authorization timestamp for each response, and a session integrity hash proving the report has not been altered. The report is designed to be filed with the court and served on opposing counsel, so every party sees the same complete record.
Case data is encrypted with AES-256 using a PIN-derived key that exists only in memory during use and is never written to storage.
Case data never leaves the device. Connectivity is used only for live speech recognition, response processing, and voice output during the hearing.
Every exchange is sealed with a SHA-256 integrity hash. Any modification after sealing is detectable.
The litigant's cloned voice is registered with cryptographic proof of ownership, restricted to system use, and revocable.
ETHER 1227 is designed to operate within established frameworks for disability accommodation in court, including Title II of the Americans with Disabilities Act, Section 504 of the Rehabilitation Act, Colorado Chief Justice Directive 04-07, the Carrie Ann Lucas Parental Rights for People with Disabilities Act, and established precedent for augmentative and alternative communication devices in legal proceedings.
ETHER 1227 answers the question every HAA engagement begins with: who holds authority over the decisions a system now touches? Here the answer is built into the system itself. The application is predominantly rule-based, the AI is limited and tightly constrained, Human Authority Architecture governs and reports on every step, and the litigant is in charge, always, by design. Where judgment must remain human, here it does.