Avi Loeb, the Harvard astrophysicist who has spent years courting both admiration and controversy for his willingness to entertain the possibility of extraterrestrial intelligence, is now turning to artificial intelligence as a tool to accelerate the search. His latest proposal: training machine learning systems to identify signatures of alien technology in astronomical data — a move that could either vindicate his unconventional approach or deepen the skepticism that has followed him for years.
The idea, which Loeb has discussed publicly and in academic circles, centers on the premise that human observers may be fundamentally limited in their ability to recognize artifacts or signals produced by civilizations far more advanced than our own. An AI system, unburdened by the cognitive biases and pattern-recognition habits of human scientists, might be better equipped to flag anomalies that would otherwise be dismissed or overlooked. As Futurism
A Scientist Who Refuses to Stay in His Lane
Loeb’s reputation precedes him. The Frank B. Baird Jr. Professor of Science at Harvard, he has authored hundreds of peer-reviewed papers and served as chair of the university’s astronomy department. But it was his 2017 claim about ‘Oumuamua — the first known interstellar object to pass through our solar system — that catapulted him into public consciousness. While most astronomers attributed the object’s unusual acceleration to natural outgassing, Loeb argued it could be a piece of alien technology, possibly a light sail. He expanded on this thesis in his 2021 book Extraterrestrial: The First Sign of Intelligent Life Beyond Earth, which became a bestseller and a lightning rod for debate.
Since then, Loeb has founded the Galileo Project at Harvard, an initiative dedicated to the systematic scientific search for evidence of extraterrestrial technological civilizations. The project has deployed instruments, analyzed materials recovered from the ocean floor near Papua New Guinea (which Loeb claimed could be fragments of an interstellar meteor), and attracted both funding and criticism in roughly equal measure. His willingness to make bold public claims before the peer-review process has fully played out has earned him the ire of some colleagues, who view his approach as more showmanship than science.
Why AI Enters the Picture Now
The application of AI to the search for extraterrestrial intelligence, or SETI, is not entirely new. Researchers at the SETI Institute and Breakthrough Listen have already been experimenting with machine learning algorithms to sift through radio telescope data for signals that might indicate intelligent origin. What distinguishes Loeb’s approach, as described in the Futurism report, is his ambition to broaden the search parameters beyond radio signals to include physical objects, optical anomalies, and other forms of evidence that traditional SETI programs have not prioritized.
Loeb has argued that focusing exclusively on radio signals — the dominant strategy since Frank Drake’s Project Ozma in 1960 — reflects an anthropocentric bias. A civilization millions of years more advanced than ours might communicate or leave traces in ways we haven’t imagined. AI, in Loeb’s framing, becomes the tool that can look without preconceptions. The machine doesn’t know what it’s “supposed” to find, and that, he contends, is precisely the advantage. Training neural networks on known natural phenomena and then flagging deviations could surface candidates for closer investigation that human astronomers would never have selected.
The Technical Challenges Are Formidable
Building an AI system capable of distinguishing between genuinely anomalous data and the far more common sources of noise, instrument error, and poorly understood natural phenomena is an enormous technical challenge. Astronomers already deal with a staggering volume of false positives. The Vera C. Rubin Observatory, expected to begin full operations soon, will generate roughly 20 terabytes of data per night. Machine learning systems trained on this data will need to be extraordinarily precise to avoid burying researchers in spurious detections.
There is also the fundamental epistemological problem: how do you train an AI to recognize something you’ve never seen before? Loeb’s proposed approach — training on known phenomena and then looking for outliers — is a well-established technique in anomaly detection, used in fields from cybersecurity to medical imaging. But the gap between “anomalous” and “alien” is vast. An unusual light curve from a distant star is far more likely to indicate an undiscovered natural process than a Dyson sphere. Critics worry that Loeb’s framework risks conflating the unknown with the extraterrestrial, a logical leap that no amount of computational power can bridge on its own.
The Galileo Project’s Growing Ambitions
The Galileo Project, which Loeb leads, has been expanding its scope steadily. The project has constructed an observatory system designed to monitor the sky for unidentified aerial phenomena — a subject that has gained renewed legitimacy following U.S. government disclosures and congressional hearings. In 2023, Loeb led an expedition to recover small metallic spherules from the Pacific Ocean floor near the path of a meteor designated IM1, which he claimed showed an interstellar origin based on U.S. Department of Defense data. He subsequently argued that the composition of these spherules was unlike any known solar system material.
However, independent analyses have cast doubt on these claims. A study published in Research Notes of the AAS suggested the spherules were more consistent with terrestrial industrial pollution — coal ash, specifically. Other researchers questioned the precision of the original fireball data used to establish IM1’s interstellar trajectory. Loeb has pushed back against these critiques, maintaining that his team’s analysis is sound and that the scientific establishment is too conservative in its willingness to consider extraordinary explanations. This pattern — bold claim, public attention, scientific pushback, defiant response — has become a recurring cycle in Loeb’s career.
Where the Scientific Establishment Stands
The broader astronomy community remains deeply divided on Loeb. Some see him as a necessary provocateur, someone willing to ask questions that more cautious researchers avoid for fear of reputational damage. Others view him as a cautionary tale about what happens when media attention outpaces the evidence. Seth Shostak, senior astronomer at the SETI Institute, has acknowledged that AI will play an increasingly important role in the search for extraterrestrial intelligence but has been careful to distinguish between the general application of machine learning to astronomy and Loeb’s specific claims about what it might find.
The tension is not really about whether AI should be used in astronomy — that question has been settled, and the answer is an emphatic yes. Machine learning is already transforming how astronomers classify galaxies, detect exoplanets, and analyze gravitational wave signals. The disagreement is about the interpretive framework Loeb places around the technology. When he describes AI as a tool that might finally detect alien artifacts, he is making a claim not just about computational capability but about the likelihood that such artifacts exist and are detectable. That is a scientific hypothesis, and many of his peers believe the evidence does not yet support the level of confidence Loeb projects.
The Broader Context: AI and the Search for Life
Loeb’s proposal arrives at a moment when the intersection of artificial intelligence and astrobiology is attracting serious institutional attention. NASA’s Astrobiology Program has been exploring how machine learning can assist in identifying biosignatures in the atmospheres of exoplanets, particularly as the James Webb Space Telescope continues to deliver spectroscopic data of unprecedented quality. The European Space Agency’s upcoming PLATO mission, designed to find Earth-like planets around Sun-like stars, will also rely heavily on automated data analysis.
In this context, Loeb’s AI ambitions are less radical than they might first appear. The difference lies in emphasis: while most astrobiologists are focused on detecting signs of microbial life — oxygen, methane, phosphine — Loeb is explicitly searching for technological signatures, or technosignatures. This is a legitimate subfield that has gained traction in recent years, with NASA funding its first technosignature research grants in 2018 after decades of congressional hostility toward anything resembling a search for intelligent aliens.
What Comes Next for Loeb and the Galileo Project
Loeb shows no signs of moderating his approach. He continues to publish prolifically, engage with the media, and expand the Galileo Project’s operations. His AI proposal, if it advances beyond the conceptual stage, would require substantial computational resources, access to large astronomical datasets, and collaboration with machine learning researchers — all of which are within reach for a Harvard-based initiative with private funding.
The question that hangs over all of this is whether Loeb’s AI-driven search will produce results that satisfy the scientific community’s standards of evidence. Extraordinary claims, as Carl Sagan famously noted, require extraordinary evidence. An AI system that flags an anomaly is only the beginning of a long process of verification, peer review, and elimination of alternative explanations. If Loeb and his team can demonstrate the rigor that process demands, they may yet shift the conversation. If not, the cycle of bold claims and scientific skepticism will continue — with artificial intelligence as its newest, most powerful, and most watchful participant.
Harvard’s Avi Loeb Wants to Build an AI That Could Detect Alien Intelligence — and the Scientific Community Is Watching Closely first appeared on Web and IT News.
