Harvard is no longer just a temple of traditional science. In the woods of Massachusetts, a group of physicists and computer scientists are redefining UFO research, one of the most controversial and fascinating fields: but don't expect bizarre conspiracy theories or fanciful tales of alien abductions. Maybe.
Il Galileo project, driven by the physicist Avi Loeb, applies the rigorous scientific method to the search for unidentified aerial phenomena, with fundamental help: artificial intelligenceInfrared cameras, acoustic sensors, and machine learning algorithms work together to catalog everything that passes above our heads, looking for that one anomaly that could change history.
Avi Loeb is certainly not a new name for those who follow the scientific exploration of the unknown. The Harvard astrophysics professor, already known for his controversial but scientifically based theories, has become a central figure in UFO research after hypothesizing in 2018 that the interstellar object 'Oumuamua could be an alien probe. A theory that earned him fierce criticism from the more conservative scientific community, but which at the same time attracted the attention of the public (and some wealthy private financiers).
Since then, Loeb has transformed what seemed like a provocation into a rigorous scientific research program. And with a decidedly innovative approach: if you want to find evidence of extraterrestrial life, don't just look for it in the depths of space, but also here, in our atmosphere, or in the ocean depths, with cutting-edge tools and a formidable new ally.
A paradigm shift in UFO research
The terminology of the discipline itself is changing, as evidence of a new scientific approach. Today, as you know, experts prefer to talk about UAP (Unidentified Aerial Phenomena – Unidentified Aerial Phenomena) rather than UFO (Unidentified Flying Objects – Oggetti Volanti Non Identificati). This is not a linguistic affectation, but an attempt to free this field of research from the sensationalistic connotations that have accompanied it for decades.
However, beyond the labels, the essence of UFO research remains unchanged: studying aerial phenomena that defy conventional explanations, with rigorous scientific methods and advanced technologies. And it is precisely this essence that is experiencing a renaissance thanks to the decisive contribution of artificial intelligence.
For decades, UFO research has been dominated by eyewitness accounts, often low-quality photos or videos, and subjective interpretations. A fertile ground for false sightings, optical illusions, and sometimes outright hoaxes. Such an approach could only generate skepticism in the mainstream scientific community.
Today, thanks to AI and increasingly sophisticated sensors, UFO research is finally acquiring the precision and objectivity necessary to be taken seriously by mainstream science. And the results may not be long in coming.
The Galileo Project Observatory
At the heart of Project Galileo is a modest but powerful observatory nestled in the woods of Massachusetts, about a half-hour drive from Boston. At first glance, it might not seem particularly impressive: a collection of cameras, acoustic sensors, radio spectrum analyzers, particle counters, weather stations, and magnetometers. But it is precisely this integrated network of sensors, linked to powerful artificial intelligence algorithms, that represents a revolution in UFO research.
Unlike traditional astronomical observatories, which focus on limited portions of the sky and objects at great distances, the Galileo Project observatory monitors the entire sky, 24 hours a day, looking for anomalies in the Earth's atmosphere. An approach that would have been impossible without the help of artificial intelligence, capable of processing the enormous amount of data generated by the various sensors.
Laura Domine, a brilliant postdoctoral physicist who chose to join Loeb’s team after studying neutrinos at Stanford, explains that AI is absolutely central to their work. “This is the only way to solve this problem,” she says with conviction. Dominé is particularly involved in developing the algorithms that filter the data collected by the observatory’s infrared cameras.
But how does AI actually work in UFO research?
The Galileo Project team is using open source computer vision software, similar to that used in self-driving cars (a recent invention could help further). These algorithms are trained to recognize common objects in the sky: airplanes, birds, insects, drones, satellites and other presences that could be confused with anomalous phenomena.
The challenge, as he explains Richard Cloete, a computer scientist at Cambridge University who oversees the software development, is particularly complex. Unlike other computer vision problems, you don't know exactly what you're looking for.
“We don’t know the characteristics of a UAP, we don’t know how it moves. This is precisely the purpose of the project: to define what a UAP is,” Cloete emphasizes.
The approach adopted is therefore that of teaching AI to recognize everything “normal” in the skies, and then identify by exclusion what does not fit into any known category. In effect, the algorithms are creating a “census of the skies,” a detailed catalog of everything that normally flies above our heads.
To train these AI models, Cloete created massive databases of real and synthetic images of common objects in the skies. Using an open-source animation program called Blender, he generated hundreds of thousands of synthetic images of airplanes, birds, drones, balloons, and other flying objects, in different lighting conditions, altitudes, and orientations.
These training sessions take place on Harvard’s compute clusters, which include hundreds of servers and specialized AI graphics processing units (GPUs), spread across three different data centers in the Boston area.
Technical difficulties and challenges
Despite the impressive computing power at their disposal, the team faces several difficulties. Planes are relatively easy for the AI to recognize, thanks to the regularity of their speed, trajectories and maneuvers. But birds pose a more complex challenge, and let alone insects (fast and erratic fliers, often close to camera lenses): they are described as a “nightmare” by Cloete.
Phenomena such as clouds, dust, leaves blowing near the horizon, and sun reflections can also confuse detection algorithms. Ironically, these same errors in perception could explain some of the mysterious UFO sightings reported by humans over the years.
The ultimate goal? Develop similar software for all the instruments in the observatory and synchronize them, so that the entire system reacts in real time to what it senses. When one of the sensors (the “all-sky” camera, the infrared array, or the microphone) detects something interesting, a special zoom camera will automatically orient itself to that point in the sky, and the entire observatory will start saving the data it is recording. The rest is legend, indeed: protocol.
First results and future prospects
A January 2025, Dominé has published a paper, co-authored with Cloete, Loeb and other collaborators, on data collected by the observatory's infrared camera array. The software YOLO (You Only Look Once) used by the team was able to identify 36% of the aircraft captured by the cameras. “By the standards of my PhD work on neutrinos, this is not a groundbreaking achievement,” Dominé admits, “but we expect significant improvements in the near future.” Yes, because this (if you haven’t figured it out) is just the beginning.
Dominé predicts that within a year or two, the Galileo observatory will be able to reliably detect anomalies from all of its sensors. In the meantime, three other similar observatories are already under development in Indiana, Nevada, and Pennsylvania.
But Project Galileo researchers aren't the only ones using artificial intelligence in UFO research. The Pentagon is also following this innovative approach with great attention.
The Pentagon's Approach to UFO Research
In 2022, the United States Department of Defense has established theAll-domain Anomaly Resolution Office (AARO), an office dedicated to the study of unidentified aerial phenomena. The creation of this office followed the sensational revelations Part 2017, when the New York Times revealed that the Pentagon had been secretly studying UFOs for years, also releasing videos taken by U.S. Navy fighter jets showing oval objects that appeared to fly in ways that defied the laws of known physics.
To date, the AARO reviewed over 1.800 UAP reports, mostly from members of the military. By cross-referencing these reports with other government data sources, such as weather and flight records, researchers determined that hundreds of cases They have simple explanations: weather balloons, clouds, drones, and so on.
However, second Jon Kosloski, current director of AARO (mathematician and engineer seconded by the National Security Agency), Some cases, about 50-60, remain “true anomalies” that have puzzled government scientists and engineers.
“There are interesting cases that I, with my background in physics and engineering and my experience in the intelligence community, don’t understand,” Kosloski told reporters in November 2024, “and I don’t know anyone else who does.”
AARO is also developing its own sensor and software system called Gremlin (Government Radar Multispectral Interrogator) to study UAPs. While most details remain classified, released information suggests the system is very similar to the Harvard Observatory: radar, radio antennas, and telescopes that collect visual and infrared images and electromagnetic radiation.
Like the Harvard team, AARO is developing custom AI software for UAP detection, and is running into similar challenges. Kosloski suggests their program is more advanced, but admits that in previous tests, AARO’s computer vision algorithm seemed to be confused by some of the details.
A global scientific collaboration
Interest in scientific UFO research is also growing outside the United States. Research programs similar to Project Galileo have sprung up at Wellesley College, the University of Würzburg in Germany, and the Nordic Institute for Theoretical Physics, as well as various government agencies.
In Europe, the MEP Francisco Guerreiro is trying to bring the instances that in the United States led to the creation of the AARO to the Pentagon, leveraging not only the safety of civil flight but also the need to provide answers to the observations of unidentified phenomena. In February 2024, Guerreiro spoke on the subject of UAPs at a plenary session in Strasbourg, highlighting the security threat posed by the lack of European reporting on the UFO issue and supporting some of the amendments he had tabled.
Italy is no stranger to this renewed scientific interest. Italian Center for UFO Studies (CISU), active for decades, is carefully following international developments. Its coordinator, Edward Russian, gave a speech to the European Parliament in March 2024, stressing how all European countries are affected by the UFO phenomenon.
Artificial Intelligence and UFO Research: Beyond Sky Observation
The application of artificial intelligence to UFO research is not limited to analyzing data from ground-based observatories. The same algorithms can be used to analyze the vast amounts of data collected by satellites, space stations, and orbiting telescopes. In addition, AI can help reexamine historical archives of sightings, photos, and videos, looking for patterns or details that have escaped human analysis.
Another promising field of application concerns the analysis of presumed materials of extraterrestrial origin. In 2021 it emerged that the Defense Intelligence Agency (DIA) of the United States had among its institutional tasks that of analyzing UFO debris. In particular, the DIA would have entrusted the study of “metamaterials” believed to be of non-terrestrial origin to the Bigelow Aerospace of Nevada. Artificial intelligence could prove invaluable in analyzing the atomic and molecular structure of such materials, comparing them to all known terrestrial materials.
The Limits of AI in UFO Research
It is also important to recognize the current problems of artificial intelligence in relation to “alien hunting”, if you will pardon the extreme simplification. As he points out Seth Shostak, astronomer at the SETI Institute and member of the advisory board of the Galileo Project, There is a fundamental difference between searching for signs of intelligent life in deep space and searching for UFOs in Earth's atmosphere.
“It’s like saying that studying unknown fauna in the rainforest is similar to those who hope to find mermaids or unicorns,” he says with a hint of skepticism.
Furthermore, AI remains a tool, however powerful, in the hands of humans, with all their beliefs, prejudices and expectations. The risk is that algorithms are more or less unconsciously programmed to find what researchers hope to find, rather than what actually exists.
Then there is the question of transparency. While the Galileo Project has committed to making its data public, many government studies, such as those by AARO, remain largely classified. This lack of transparency inevitably fuels conspiracy theories and undermines the scientific credibility of the research.
The Future of UFO Research in the Age of Artificial Intelligence
While there are, as you have read, problems to be solved, the future of UFO research looks more promising than ever, and thanks to artificial intelligence. AI's ability to analyze huge amounts of data, identify complex patterns, and continuously learn makes it the ideal tool to tackle one of humanity's greatest mysteries.
In the coming years I expect significant advances in both sensor technology and artificial intelligence algorithms dedicated to UFO research. The integration of different data sources (ground observatories, satellites, military and civilian radar) could create a global monitoring network capable of detecting and analyzing any anomaly in our skies.
Growing interest from prestigious scientific institutions and government agencies will move UFO research from the realm of pseudoscience to that of rigorous science.
Towards a new scientific paradigm
The real revolution in UFO research in the age of artificial intelligence is not only technological, but also cultural and epistemological. For the first time, we are addressing the possibility of the existence of non-human intelligences with the tools of modern science, rather than folklore or speculation.
As emphasized Avi Loeb, “the universe is a very strange place and we cannot be dogmatic when most of the things we discuss are not understood.” This open-mindedness, combined with scientific rigor and the potential of artificial intelligence, could finally lead us to answer the question we have been asking for millennia: are we alone in the universe?
Whether it's alien probes, as yet unknown natural phenomena, or secret terrestrial technologies, the scientific approach promoted by projects like the one at Harvard and the Pentagon will allow us to shed light on one of the most fascinating mysteries of our time. And perhaps, one day not too far away, AI will help us find that definitive proof that will forever change our view of the cosmos.
