The Hall Lab.
The Hall Lab studies what an autonomous system has to demonstrate before it is allowed to operate, and who has to be trained to judge it. Aviation safety, aviation security, and the operational management of a real flying organisation, brought to bear on Physical AI.
Led by Tasjah Hall, MS, aviation consultant and CEO of Limitless Altitude, Inc.
Four areas.
Each is drawn from the lead's operating record rather than from a research agenda written in advance.
What a system must show before it flies
Aviation safety practice applied to autonomous systems: the evidence an operator has to produce, the failure cases that have to be covered, and who signs.
Operations under an adversary
Aviation security as an operational discipline rather than a software property: access, screening, and the procedures that hold when a system is targeted.
Management of a real flying organisation
Twenty-five years of aviation management and operations: scheduling, compliance, crew and ground coordination, and what automation actually changes about each.
The people qualified to judge autonomy
Aviation instruction at scale: what a curriculum has to contain before a graduate can be trusted to certify, operate, or refuse an autonomous system.
Tasjah Hall, MS
Lead · The Hall Lab · Academic Dean · Program Coordinator
Ms. Hall is an aviation consultant and the CEO of Limitless Altitude, Inc., with more than 25 years of aviation management and operations experience, including aviation safety and security. She served as an Aviation Instructor at Texas Southern University for over 16 years. As Academic Dean she carries the Institute's academic programs and their coordination.
What is missing, and what would move it.
This lab has not published yet. Rather than list work it has not done, here is what it is for, in the form the Institute uses for every open problem.
| What is missing | Why | What builds it | What becomes possible |
|---|---|---|---|
| A statement of what an autonomous system must demonstrate before it operates | Aviation has a mature answer for crewed flight and a partial one for small UAS. Physical AI outside aviation largely has none, and each programme invents its own. | Mapping the existing aviation safety case onto an autonomous system, and marking which parts transfer, which do not, and which have no counterpart yet. | An operator can be told, in advance, what evidence will be asked of them, instead of discovering it at the point of refusal. |
| A curriculum that qualifies someone to judge an autonomous system | Training produces people who can fly, build, or program. Judging whether a system is fit to operate is a separate competence and is taught almost nowhere. | Sixteen years of aviation instruction turned into a competency ladder with an assessment at each rung. | A workforce that can certify and, when required, refuse. Refusal is the part that cannot be automated away. |
| Aviation security treated as an operational discipline in Physical AI | Security in autonomy is usually scoped to the software. An aviation operator scopes it to the whole organisation: access, screening, procedure, and the people running them. | The operational security practice of a flying organisation, written down against the failure modes an autonomous fleet actually presents. | A security posture that survives contact with an adversary who is not attacking the model. |