The green line on the map Jon Kosloski showed U.S. senators on November 19, 2024, traced the path of the airplane carrying the camera. It had flown in an arc around an airport in Puerto Rico. In its infrared footage, a small shape appeared to race past buildings, split in two and disappear into the Atlantic. On Kosloski’s map, the objects never made it past the airport.

Kosloski is an electrical engineer and the director of the Pentagon’s All-domain Anomaly Resolution Office, or AARO. The office investigates reports of unidentified anomalous phenomena, the broader term the government uses for what most people still call UFOs. Kosloski had spent more than two decades at the National Security Agency before he took the job. Asked to explain a famous sighting in public, he started with where the observer had been.

What makes him worth listening to is the distinction he drew next. His team was confident the objects had stayed in the air. He was less certain what they were. Later, when six federal agents described lights above a ridgeline in the American West, he signed a report that left part of their account open. What allowed him to answer a question about a stretch of ocean eleven years after the fact, yet decline to identify lights seen by three teams of trained observers?

Official portrait of Jon Kosloski in a blue suit, with the U.S. flag behind him.
Jon T. Kosloski studied how to detect faint optical signals before becoming director of the Pentagon’s All-domain Anomaly Resolution Office in August 2024. Credit: AARO / U.S. Department of Defense, via Wikimedia Commons

Before UFOs, a much fainter signal

Kosloski earned degrees in mathematics and physics at California State University, San Bernardino, and a doctorate in electrical engineering at Johns Hopkins. His Pentagon biography describes doctoral work on devices for receiving extremely weak optical signals. At the NSA’s Research Directorate, his work ranged from optics to computing and the mathematics of codes.

One 2007 paper he coauthored gives that résumé a little texture. The researchers wanted to know whether information carried by individual particles of light could travel through an optical fiber alongside much brighter, ordinary communications traffic. It could, under the conditions they tested. But light from the brighter channels scattered inside the fiber and interfered with the faint signal. Knowing what went into the line was only half the experiment. They also had to measure what the line itself contributed to what came out.

An experiment like that can be run again. Nobody can send an aircraft back to Puerto Rico on the night of April 26, 2013. When the Pentagon appointed Kosloski in August 2024, he inherited an office established in 2022, first led by Sean Kirkpatrick and then by acting director Tim Phillips. His brief was to help determine what had been observed near sites of national security interest, whether it posed a risk, and what the government might have missed. Often he would have a recording made for some other purpose, with no way to ask the camera to take a second look.

Where did the light go?

The original Aguadilla video came from a U.S. Customs and Border Protection aircraft near Rafael Hernández Airport. Watch the background: land gives way to ocean, and the dark shape grows hard to see. It is easy to read the sequence as a splashless dive. But a camera that sees the ocean behind an object hasn’t established that the object is above the water. From the right angle, you can photograph a bird against the sea while it is still over the beach.

At the Senate hearing, Kosloski explained what the other lines on his map were for. They showed where the camera had been pointing at different moments as the plane turned. His team used the aircraft’s changing position and those viewing directions to reconstruct a path for what it filmed. In that reconstruction, two small objects drifted over land at about eight miles an hour, roughly with the wind. The camera’s own motion made the background sweep past them. And because it recorded differences in infrared brightness, a warm object could fade from view as its contrast against the background diminished. Disappearing from the picture did not require entering the water.

The written case assessment, dated March 20, 2025, attached different levels of confidence to two conclusions. AARO had high confidence that the objects showed no extraordinary performance or passage between air and water. It had moderate confidence that they were a pair of sky lanterns. Local vendors told investigators that resorts released lanterns at celebrations, and lanterns fit the estimated size and fading heat signature. That establishes a plausible local source. It doesn’t identify a particular celebration or the people who released them. In an August 2025 conversation on StarTalk, Kosloski still allowed for balloons. His confidence about the objects’ path didn’t oblige him to supply a name for them.

A small dark shape appears just right of the tracking marks in the original Aguadilla infrared recording.
A dark shape appears just right of the crosshairs in this infrared recording made near Aguadilla, Puerto Rico, on April 26, 2013. The frame shows where the crew’s camera was looking; by itself, it does not establish how far away the shape was. Credit: U.S. Customs and Border Protection; released by AARO via DVIDS

The people who asked for the coordinates

The Scientific Coalition for UAP Studies, a civilian research group, had reached a different conclusion before AARO existed. Its 2015 analysis estimated speeds approaching 100 miles an hour over land and interpreted the final sequence as an object entering and leaving the ocean. On April 28, 2025, the group published the questions it had sent AARO. It asked for the reconstructed positions and their margins of error, the weather observations placing the aircraft in scattered cloud, and the analysis of a few frames in which the object seemed to pass behind a utility pole.

That last question may sound small. It matters because a pole is something you can locate on the ground. If it briefly blocked the object, it could help establish how far away the object was. An apparent overlap that never actually hid the object would tell you much less. AARO said it had examined the pixels and found no passage behind the pole. The group wanted to see how it reached that judgment. It said the office had acknowledged its questions but was not prepared to answer them at the time. That exchange identifies a real limit on independent checking. It supplies no new evidence of a dive into the Atlantic.

Kosloski has given a reason the gap can be hard to close. In a December 2025 interview with the Naval Postgraduate School, published in its spring 2026 CTX issue, he explained that AARO does not own many of the systems from which it gets data. Releasing material can require the cooperation of the service that collected it. He also described a less dramatic limit: sometimes a blurry public image really is as good as the original gets. Those are two different problems. Declassification can help with one. It can’t bring a distant point of light into focus after the fact. Neither observation tells us which specific answers, in the Puerto Rico case, the office could have released.

AARO’s overhead map of the airport and aircraft route beside the corresponding infrared frame.
AARO’s reconstruction pairs an overhead map of Aguadilla with the aircraft’s infrared view. Analysts placed the objects’ path over land. They were more confident that the objects never entered the water than that they were sky lanterns. Credit: AARO, via DVIDS.

Six witnesses and a missing distance

In October 2023, three pairs of federal law-enforcement agents reported lights near a sensitive site in the western United States. Across two days, their accounts described a larger orange light that seemed to release smaller red ones. One red light reportedly held its place above a ridgeline for hours. In a June 5, 2026 memorandum, Kosloski put his name to an assessment that took the similarities in the accounts seriously. The agents knew the area, and they had watched from different positions. They had not collected imagery or measurements of the lights themselves.

Asked on StarTalk how he approaches witnesses, Kosloski said he starts by assuming people are sincere. “And then we look to corroborate that.” Here the agents’ descriptions gave analysts times and directions they could check against flight records. Military aircraft had been releasing flares, and their activity plausibly accounted for roughly 60 percent of what was reported. The match did not hold for the rest. In one account, for instance, the light reportedly stayed in place longer than a flare would burn. The team could establish that flares had been used in the area. It could not establish how far that particular light was from the agents, an elementary fact that would also bear on why they heard no sound.

Kosloski’s memo retained “unrecognized technology” as a hypothesis for some of the remaining observations. It explicitly said that possibility rested on testimony and the elimination of alternatives, without technical or physical evidence to substantiate it. The percentage in the memo describes how much of the reported activity a proposed explanation might cover. It is not a probability that an unknown technology was present. Keeping the case open meant investigators had a reason to look again. It did not mean they had detected a new kind of machine.

The Reality Signal

Kosloski’s work exposes a habit in how we talk about UFOs. We give the object a label, unidentified, when the uncertainty may belong to just one part of a sighting. Over Aguadilla, AARO could make a strong argument about where the objects traveled and only a tentative one about what kept them aloft. In the western case, the investigators knew something solid about aircraft in the area and had consistent descriptions from witnesses, yet lacked the distance to a light that mattered. Calling both stories “unexplained” would throw away much of what we have learned.

Why It Still Matters

AARO’s fiscal 2025 annual report, made public in July 2026, puts that distinction to work. Of 319 incoming reports, 191 went into an archive where they could be reopened if useful evidence arrived; the office separately flagged nine reports for further analysis. A file waiting for another observation and one promising enough to warrant investigators’ time are both open. They make very different demands on the next person who picks them up.

What surprised Kosloski, he told Walzer, was how often an intriguing case still lacked a useful measurement in a world so full of sensors. Asked about a possible extraterrestrial origin, he said AARO had found no verifiable evidence for one in the cases it had investigated. That is a statement about the evidence of visits to Earth. The existence of life elsewhere is a different question.

Two excerpts from Kosloski’s memo show the hypothesis being considered and the lack of technical or physical evidence to substantiate it.
Flight records could plausibly explain some of the lights six federal agents reported in the western United States in 2023. In his June 5, 2026 memo, Kosloski left the remaining observations unresolved. The witnesses’ accounts did not establish that a new technology was present.

The Galactic Mind Perspective

Kosloski is most persuasive when he gives someone who disagrees with him something specific to test. The moving aircraft, the shoreline and the fading thermal contrast make a stronger account of Aguadilla than the impression that a shape dove beneath the waves. We would give that explanation real weight and leave the lantern identification at the more modest confidence his office assigned it. We would also publish the reconstructed positions and their uncertainties wherever they can be released. A diagram can make an argument intelligible without giving another analyst enough to check the arithmetic.

There is a human obligation on the other side of that exchange. At the 2024 hearing, Senator Kirsten Gillibrand asked what the pilots who filed reports heard back. Kosloski acknowledged that feedback was informal and described a system he wanted to build. It was a good question to put to him. A witness who learns which detail made a difference in one investigation is in a better position to notice it the next time. That may be one of the most useful things a UFO office can return to someone who looks up and sees something strange.

Open Thread

If two investigators put the same light on opposite sides of a shoreline, what would a third person need to decide between their maps?

More in Dossier

Sources / Receipts