Artificial intelligence is currently eating the world. Whether you believe it’ll lead to superintelligent machines, the end of humanity, or a bursting bubble, its impact can already be felt in industries around the globe. 

It’s also spreading beyond our little blue and green planet. We’ve got AI in satellites, IBM’s Watson did a tour aboard the International Space Station, and NASA is currently working to integrate AI technologies into future spacecraft.

Soon, if a pair of scientists at the forefront of the search for alien life get their way, we’ll be sending chatbots like OpenAI’s ChatGPT to explain humanity and life on Earth to extraterrestrials.

The search for ET

Franck Marchis, director of citizen science at the SETI Institute, and NASA principle research engineer Ignacio Lopez-Francos recently penned an op-ed for Scientific American describing the potential risks and rewards associated with sending an AI to communicate with any intelligent life that finds it.

The big idea here isn’t the creation of an android or sentient computer capable of speaking on our behalf. It’s more like putting an interactive historical record on a disk.

Back in 1977, when the United States launched the Voyager 1 spacecraft, scientists put a 12-inch, gold-plated, copper disc onboard. The record contained data including pictures, music and other sounds, and texts describing our planet and its diversity.

US president Jimmy Carter encoded the disc with a message for any aliens who might find the spacecraft:

"This is a present from a small distant world, a token of our sounds, our science, our images, our music, our thoughts, and our feelings. We are attempting to survive our time so we may live into yours.”

In 1998 the Voyager 1 spacecraft passed the Pioneer 10 craft to become the farthest human-made object from Earth. It was about 15 billion miles away in 2023. So far, nobody’s answered the call.

Modern communications

Now that our technology has evolved, Marchis and Lopez-Francos believe it’s time to update the message. Rather than sending snippets of humanity in the form of images, songs, and writings, they posit that an LLM could allow aliens to ask their own questions and probe the depths of our recorded knowledge.

Per the scientists:

“This would enable extraterrestrial civilizations to indirectly converse with us and learn about us without being hindered by the vast distances of space and its corresponding human lifetime delays in communication.”

They go on to point out that there are risks involved. Hostile aliens could, for example, use that knowledge against us. But, moreover, we’d need to figure out how to send aliens a usable AI system capable of operating without the internet.

We could take the Voyager 1 route and put the tech on a drive then send it off on a collision course with destiny. But, it’s taken Voyager 1 nearly 50 years to get a meager 15 billion miles away from its home. At this rate it would take the spacecraft until somewhere around the year 3084 to reach our nearest galactic neighbor, Alpha Centauri.

The alternative would be to send a signal into space with the data necessary to transmit an LLM to anything in the universe capable of using it. The challenge then lies in our ability to transmit data across vast distances at high speed.

NASA currently has installations near the Earth’s moon capable of transmissions in the 100 megabytes per second range. This is somewhat on par with what we can do terrestrially. But, as the scientists put it, “interstellar communication with current technology would likely drop to 100 bits per second.”

This means it would take centuries to send a modest-sized model such as Meta’s Llama-3-70B as far as the next galaxy over. With certain data-shrinking techniques however, the scientists believe that time could be reduced to about 20 years or so.

While this all remains hypothetical, and there’s plenty of technical details to sort out, there’s at least one non-scientific question that would need to be sorted far ahead of time: what AI model should we send?