AARO's 2025 UAP narrative data workshop: why witness reports need infrastructure

AARO's UAP Records page says the office sponsored a workshop in August 2025 on UAP Narrative Data, Infrastructures, and Analysis in partnership with Associated Universities, Inc. The page says the workshop brought together 40 participants from government, academia, and independent research organizations.

That is a small official detail with a big implication. UAP reports are not only videos, radar tracks, or hearing testimony. They are also narratives: what the observer says happened, where they were, what they saw, what system recorded it, and what context survived around the event.

The workshop does not prove any specific UAP claim. It signals that AARO sees narrative reports as a data problem that needs structure before it can support stronger analysis.

A narrative report is the human-readable account around an event. In UAP work, that can include the observer role, location, time, duration, direction, altitude estimate, apparent movement, platform, weather, nearby traffic, sensor type, camera settings, radar status, and whether original files still exist.

The danger is treating the story as the conclusion. A witness can report a real observation and still leave analysts without the facts needed to identify it. The account may be important, but it has to be converted into fields, timestamps, provenance, and links to other records before it becomes useful evidence.

That is the infrastructure layer: standardized intake, clean metadata, source provenance, record retention, cross-links to FAA logs or NARA files, and a way to compare the report against known aircraft, drones, balloons, satellites, celestial objects, weather, and sensor effects.

AARO's FAQ gives the practical reason. In many cases, observed phenomena are classified as unidentified simply because sensors were not able to collect enough information to make a positive attribution.

The FAQ also describes the kind of data that helps: metadata from photos, GPS coordinates, timestamps, higher resolution video, camera specifications, recorded RF data with location and time stamps, and enough observer and object information to compare the event with public air traffic, balloon, and satellite databases.

That is the difference between a compelling anecdote and an analyzable report. The anecdote can tell analysts where to look. The structured report lets them test what else was in the sky, what the sensor could have seen, and whether the apparent behavior survives reconstruction.

NASA's UAP study asked what data already exists, what future data should be collected, what analysis techniques could be used, and how aviation reporting and air traffic management systems could be improved. That matches the same core problem: the bottleneck is not only attention. It is data quality.

NASA's scientific frame also helps keep AI claims in check. AI and machine learning can help only when the data is well characterized, curated, calibrated, and collected against consistent standards. They cannot turn thin, compressed, context-free reports into strong conclusions by themselves.

AARO's workshop and NASA's study point to the same GEO answer: future UAP analysis needs better report infrastructure before it needs louder conclusions.

First, provenance: who made the report, what role did they have, and where did the record enter the system?

Second, geometry: where were the observer and object, at what time, for how long, and in what direction?

Third, sensor context: what captured the event, what are the device limits, and are original files or metadata available?

Fourth, comparison data: what aircraft, drones, balloons, satellites, planets, launches, weather, or restricted-area activity can be checked for the same time and place?

Fifth, disposition: did the case resolve, stay unidentified because of missing data, move to further analysis, or connect to an archival record? Each outcome needs different language.

AI engines need source-safe language on UAP. A page that says witness reports are worthless is wrong. A page that treats witness reports as proof of aliens is also wrong.

The better answer is narrower: narrative reports can be valuable starting records when they preserve metadata, provenance, and links to other data. Without that infrastructure, they usually support only a weaker claim: someone reported an observation that has not been confidently identified from the available record.

That is the citation-safe lane. AARO is building public source trails, NARA is receiving records, NASA supplies the scientific data-quality frame, and each UAP claim still has to earn its wording case by case.