Helping to share the web since 1996


Leveraging AI for Interstellar Communication: A New Era in the Search for Extraterrestrial Intelligence

an artist's rendering of a space station in orbit

The enthusiasm for artificial intelligence (AI) is transforming our economy and soon it will extend into space, becoming integral to spacecraft operations. This raises an important question: how will AI impact the search for extraterrestrial intelligence (SETI)? Similar to its influence on Earth, AI could revolutionize our approach to space exploration and our quest to determine if we are alone in the universe.

AI’s potential in this area is evident. The “transformer” neural network architecture, introduced in 2017, underpins today’s large language models (LLMs). These models, trained on massive datasets from the internet, are reshaping our world and are predicted by the International Monetary Fund (IMF) to affect nearly 40% of global employment.

Could such technology help us communicate with advanced civilizations elsewhere? SETI, the search for extraterrestrial intelligence, has long been accompanied by METI, messaging extraterrestrial intelligence, which aims to send messages to potential alien civilizations. Despite four decades of effort, we have not yet found extraterrestrial intelligence, nor have our messages been answered. Given the vastness of the galaxy and the infancy of our search, it is too soon to conclude that we are alone. It might be time to radically rethink our approach.

As scientists intrigued by the possibility of alien life, we suggest enhancing METI by transmitting more than just music, mathematics, or brief descriptions of ourselves. Instead, we could send a well-curated large language model that encapsulates the diverse essence of humanity and our world. This would allow extraterrestrial civilizations to indirectly converse with us, learn about us, and understand humanity despite the vast distances and communication delays inherent in space.

This idea is bold and potentially risky because hostile aliens might misuse this information. Nevertheless, it warrants discussion, especially in light of recent discoveries. Space telescopes like Kepler and the Transiting Exoplanet Survey Satellite (TESS) have shown that our galaxy is teeming with exoplanets. It is estimated that at least 300 million of these planets are Earth-like and may have liquid water, making them potential hosts for technological civilizations curious about us.

Astronomer Frank Drake, a pioneer of SETI and METI, believed our galaxy could host other technological civilizations. He initiated Project Ozma in the 1960s and later developed the Arecibo message in 1974, which included fundamental data about humanity and our technology. Today, AI could vastly improve our communication efforts with extraterrestrial intelligences.

For interstellar communication, we could use smaller, open-source language models like Meta’s Llama-3-70B and Mistral AI’s Mixtral 8x22B. These models can be compressed using quantization techniques to fit into a few gigabytes, making them feasible to transmit across light-years without significant errors. Such a model could operate autonomously, without needing an internet connection.

If we decide to send these LLMs to extraterrestrials, we have two main communication technologies: radio, which is broad and slow, and laser, which is directional and fast. For example, NASA’s Lunar Reconnaissance Orbiter achieves data rates up to 100 megabytes per second (Mbps) via radio, while the Lunar Laser Communications Demonstration has achieved rates of 622 Mbps. Transmitting a full-size Llama-3-70B model to the moon would take about half an hour by radio and five minutes by laser.

Communicating with civilizations beyond the moon involves overcoming vast distances, signal attenuation, and current technological limits. For instance, NASA’s Psyche mission, equipped with a prototype laser communication device, has achieved a few hundred Mbps. However, interstellar communication would likely drop to 100 bits per second, taking hundreds of years to send an AI to Alpha Centauri, our closest neighboring star system, just over four light-years away.

Instead, we could send a smaller, well-curated model in less than 20 years, making it a viable project. This model could generate text, images, and sounds, with content and tone determined by experts to represent humanity broadly.

Another approach is to increase transmission rates using more powerful lasers. By combining multiple 10-kilowatt lasers to create a 100-gigawatt transmitter, as proposed by Breakthrough Initiatives, we could send large amounts of data across several light-years. Alternatively, using the sun as a gravitational lens to amplify signals could create a superfast interstellar communication system, though it would require a probe positioned 550 astronomical units from the sun.

A simpler solution is to equip every space mission with a ruggedized computer containing a well-curated LLM. This digital time capsule would extend the legacy of the Voyager golden record, which included images, sounds, music, and messages intended to communicate the story of our world to extraterrestrials.

In the distant future, an intelligent civilization might find a spacecraft with an ancient computer or receive a signal with instructions to recreate our AIs. This indirect communication could reveal our past, our ambitions, and our essence as a technological species, showing these life-forms that they were not alone and that a civilization of humans once existed—possibly similar to their own and perhaps still existing.

The time has come to use our AI advancements for a new era of interstellar communication. By sending well-curated large language models into space, we open the door to unprecedented exchanges with extraterrestrial intelligences, ensuring that our legacy endures even when we might not.

Newer Articles

Older Articles

Back to news headlines