🛸 If you could talk to aliens, what would you say? What once seemed like pure science fiction is now being seriously discussed by scientists worldwide. Thanks to advances in AI technology, researchers are proposing revolutionary ways to communicate with extraterrestrial intelligence.
When AI Becomes Humanity's Ambassador
In July 2024, Dr. Franck Marchis of the SETI Institute and Ignacio G. López-Francos of NASA's Ames Research Center published a groundbreaking proposal in Scientific American. Their idea: send large language models (LLMs) into space to enable "indirect conversations" with extraterrestrial civilizations.
Traditional approaches, like the 1974 Arecibo Message, transmitted basic information using simple binary code—numbers, DNA structure, the human form—comprising just 1,679 bits of data. While pioneering, this method could barely scratch the surface of human knowledge and culture.
The new proposal turns this approach on its head. By transmitting an LLM trained on humanity's collective knowledge, any receiving alien civilization could learn our languages, ask the AI questions, and develop a deep understanding of who we are as a species.
Technical Challenges and Solutions
This ambitious plan faces several technical hurdles.
First, there's the data size problem. Meta's Llama-3-70B weighs in at approximately 130GB—far too large for interstellar transmission. However, using a technique called "quantization," researchers can compress the model to just a few gigabytes while maintaining performance.
Then there's transmission speed. While NASA's lunar communication can reach 100 Mbps, interstellar signals drop to roughly 100 bits per second. Even with compression, sending a model to Alpha Centauri (our nearest stellar neighbor at 4 light-years) would take nearly 20 years.
Researchers are exploring faster transmission technologies, including the 100-gigawatt laser transmitter proposed by Breakthrough Initiatives and the "Solar Gravitational Lens Mission" that would use the Sun as a natural signal amplifier.
AI Revolutionizes SETI on Both Ends
AI isn't just transforming how we might send messages—it's revolutionizing how we search for extraterrestrial signals.
At GTC 2025, SETI Institute's Luigi Cruz presented a novel method using NVIDIA GPUs and neural networks. His team synchronized 42 antennas at the Allen Telescope Array to identify radio pulses from a pulsar in the Crab Nebula in real-time.
This wasn't just a tech demo. It proved that AI can effectively find "needles" of potential alien signals in the cosmic "haystack" of space noise.
In Japan, space venture SpaceShift announced in 2024 plans to apply their satellite data analysis AI expertise to SETI research. Working with Professor Keitaro Takahashi of Kumamoto University, they're analyzing radio telescope data from observatories worldwide.
Updated Post-Detection Protocols
What happens if we actually detect an alien signal?
In October 2025, the International Academy of Astronautics' SETI Committee released updated Post-Detection Protocols—the first revision in 15 years. Led by Professor Michael Garrett, the new guidelines address modern realities like social media influence and AI technology advances.
The core principles remain:
- Thoroughly verify any detected signal
- Don't announce until confirmed
- Once confirmed, disclose openly without concealment
- Don't reply without international consultation
That last point is crucial. As Stephen Hawking warned, contact with a more technologically advanced civilization carries inherent risks. Even the LLM transmission proponents acknowledge this is "a radical and potentially risky idea" while arguing the discussion is worth having.
New Search Strategies
A 2025 study from Penn State University and NASA's Jet Propulsion Laboratory proposed an intriguing new approach: assume alien civilizations communicate with their spacecraft just like we do.
NASA's Deep Space Network regularly beams signals to planets like Mars. These transmissions "spill over" along the orbital plane and could theoretically be detected from outside our solar system. The research team proposes that alien civilizations likely do the same—and we should look for their "leaked" signals during planetary alignments.
Using this method, signals from civilizations within 23 light-years could potentially be detected. Observations of the TRAPPIST-1 system using this technique have already begun.
Updating Voyager's Legacy
In 1977, the Voyager spacecraft launched carrying a "Golden Record"—an analog time capsule containing music, images, and greetings in 55 languages representing human culture.
The LLM transmission proposal essentially updates this legacy for the digital age. Future space missions might carry ruggedized computers loaded with LLMs that can operate without internet connections.
Someday, a distant civilization might discover one of these spacecraft, boot up its AI, and begin learning about humanity—our languages, history, and culture. Even if humans no longer exist by then, our knowledge and memory would live on somewhere in the cosmos.
What About Your Country?
In Japan, researcher Shinya Narusawa at Nishiharima Astronomical Observatory has led SETI research and coordinated international observation projects. Now that the romance of searching for extraterrestrial life meets cutting-edge AI technology, what do you think?
How does your country approach the search for extraterrestrial intelligence and AI-powered space research? If you could send a message to aliens, what would you want to say? Share your thoughts in the comments!
References
- https://www.scientificamerican.com/article/artificial-intelligence-will-let-humanity-talk-to-alien-civilizations/
- https://developer.nvidia.com/blog/how-seti-uses-ai-to-search-for-intelligent-alien-life/
- https://astrobiology.com/2025/10/seti-post-detection-protocols-progress-towards-a-new-version.html
- https://www.astronomy.com/science/to-find-alien-life-try-eavesdropping-on-their-conversations-with-spacecraft-astronomers-suggest/
- https://uchubiz.com/article/new43696/
Reactions in Japan
The idea of sending AI into space is amazing. It's like a modern update of Voyager's Golden Record. But sending 130GB of data over decades... that's mind-boggling.
I was moved when I learned about Dr. Narusawa's research at Nishiharima Observatory during college. Japan has serious SETI researchers too. The fusion with AI might mark a new era.
Honestly, there's no guarantee aliens could understand LLMs even if they exist. The definition of 'intelligence' in intelligent life seems too human-centric.
This is literally sci-fi becoming reality. Makes me want to reread Carl Sagan's 'Contact'. An AI ambassador for humanity traveling through space could make a great novel.
We should take Hawking's warning seriously. Packing all human information into an LLM means exposing all our weaknesses too. Terrifying from a cybersecurity perspective.
Even after humanity is gone, AI would continue sharing our memories somewhere in the cosmos... that's moving. An eternal footprint of humanity.
Watching SpaceShift's efforts closely. Applying satellite data analysis to SETI might be uniquely Japanese thinking. Feeling the potential of space business × AI.
The method of targeting planetary occultations is really clever. Disappointed they found nothing in TRAPPIST-1 though. Still worth searching, I think.
Who decides what goes into 'AI representing humanity'? Worried it'll be Western-centric. Questionable if it can properly reflect diverse cultures and values.
The story about detecting pulsar signals with NVIDIA GPUs is quietly amazing. If AI can distinguish signals from noise, SETI research could accelerate rapidly.
20 years to send to Alpha Centauri, earliest reply in 8 more years? Won't see results in my lifetime. Have to hope for the next generation.
The idea of compressing LLMs with quantization is interesting. But how much 'humanness' a compressed model can convey is unknown.
If we contact extraterrestrial intelligence, humanity's religious and worldviews would be fundamentally shaken. I doubt many are prepared for that.
When I tell visitors 'We now use AI to search for aliens', kids' eyes light up. Hope it sparks their interest in science.
I get the argument that SETI budget should go to Earth's problems instead. But I think humanity needs research that satisfies intellectual curiosity too.
As a SETI scientist, I'm excited about this proposal. LLMs could enable much richer information exchange than simple beeps. However, sending without international consensus might be premature.
Interest in space research is growing in Sweden. The idea of using AI for alien communication is innovative, but I think we need more ethical discussions first.
India's ISRO is also interested in SETI-related research. AI-powered signal detection could enable effective research even with limited budgets.
The French astronomy community is paying attention to this. But who decides the content of an AI representing humanity, and how do we ensure cultural diversity?
From Ireland here. Sending LLMs to aliens sounds like a sci-fi movie! But seriously, this could become the biggest project in human history.
From a German perspective, this project needs an international legal framework. From the Outer Space Treaty standpoint, no single nation or institution should act unilaterally.
Astronomy fan from Brazil. I think observations from the Southern Hemisphere are important too. Combined with AI, we might discover signals we've been missing.
Russia has a long history of SETI research too. I see potential in AI-based new approaches, but international cooperation rather than competition is essential.
Working at Australia's Parkes Observatory. Our radio telescope could be used for AI-assisted searches. Adding Southern Hemisphere perspective greatly expands search coverage.
Korean space scientist here. This proposal is very interesting, but there's no guarantee aliens can understand LLM languages. We should also consider more universal communication methods.
Living near the former Arecibo Observatory in Puerto Rico. 50 years since the 1974 Arecibo Message, technological progress is remarkable. It's moving that AI inherits that legacy.
From Nigeria. Concerned that African perspectives aren't included in this discussion. If creating 'AI representing humanity', all continents' voices should be reflected.
Canadian space enthusiast. I personally support SETI research, but taxpayer-funded projects need careful discussion. AI might enable efficient, low-cost research.
China's FAST telescope can also contribute to SETI research. Combined with AI technology, we can maximize the capabilities of the world's largest single-dish telescope.
Reading this from Italy. The Vatican Observatory is also interested in this field. Discovering aliens might prompt reconsideration of the relationship between religion and science.