“Project Hail Mary,” a science fiction novel that has just been adapted into a big-screen motion picture, tells the story of an unlikely astronaut who unexpectedly encounters an alien during a desperate mission to save their respective civilizations.
The astronaut (played by Ryan Gosling in the film) and the alien must immediately find out whether they are friends or enemies. You also need to develop a translation system that supports two completely different types of communication.
It all makes for a life-and-death space drama reminiscent of “Apollo 13” – but the day is fast approaching when advances in astronomy and artificial intelligence could largely remove the drama from extraterrestrial contact.
Seth Shostak, senior astronomer at the SETI Institute, says he wouldn’t be at all surprised if our first encounter with aliens came in the form of AI-to-AI contact.
“I suspect the aliens will be machines, because that’s what we do, right?” he says in the latest episode of the Fiction Science podcast. “We are in the early stages of building machines that can do things that humans have had to do in the past. I’m sure that in 100 years the most powerful intelligence on this planet will no longer be a soft and soft biological thing. That will be a machine. So if we hear the aliens, I suspect it’s more than likely that they too will be machines.”
If you’re worried that discussing AI and the search for aliens requires a deep dive into spoilers, have no fear: Artificial intelligence plays no real role in the movie “Project Hail Mary.” In the novel by Andy Weir on which the film is based, it is mentioned only once – only to explain why the planners of the “Life or Death” mission decided against using AI. (However, we will encounter spoilers towards the end of this post, so be warned.)
For more than 65 years, astronomers have searched the sky for radio signals that may have been emitted by extraterrestrial civilizations. “The usual approach is to build a receiver that can monitor thousands – today even millions – of different channels at the same time,” says Shostak. “And you can just look at how this ability has improved over time. It turns out that it follows something called Moore’s Law… which says that the speed of electronics more or less doubles every two years.”
It takes a lot of computing power to monitor millions of channels, and Shostak is sure that AI will accelerate the search for extraterrestrial intelligence, better known as SETI.
There’s already evidence of this: Last November, the Breakthrough Listen Initiative reported that an AI system developed in collaboration with NVIDIA could process real-time data from telescopes searching for fast radio bursts at a speed more than 600 times faster than the current data pipeline. The system improved detection accuracy by 7% and reduced false alarms by almost an order of magnitude.
“This technology not only allows us to find known signal types more quickly – it also allows us to discover completely unexpected signal morphologies,” said Andrew Siemion, principal investigator of the Breakthrough Listen Initiative, in a press release. “An advanced civilization could use burst communications, modulated signals, or transmission schemes that we haven’t even imagined yet. This AI system can learn to recognize patterns that might be completely missed by a human.”
A few years ago, another team of astronomers used a machine speech algorithm to identify potential extraterrestrial signals missed by other data processing systems. (But don’t get too excited: Follow-up observations did not confirm that the signals came from alien civilizations. They would have heard it if they did.)
AI tools could help astronomers overcome some of the obstacles facing the SETI search. For example, a group of researchers recently reported that signals from extraterrestrial civilizations could be disrupted by stormy space weather. Improved pattern recognition software could potentially detect the signal hidden in the cosmic noise.
AI models could also come into play to interpret alien messages as they are found. But Shostak isn’t so focused on that challenge. “Even if we never understand what the aliens are saying, just the fact that we can receive the signal and recognize that it is an artificial signal – in other words, generated by some technology – is very interesting, because we have proven that they are there,” says Shostak.
Understanding what the aliens are saying “would be interesting to know, but I would consider that a secondary benefit to detecting their presence,” he says.
Seth Shostak is a senior astronomer at the SETI Institute. (Photo by SETI Institute)
Shostak compares the challenge of deciphering alien messages to the challenge archaeologists faced when excavating Egyptian hieroglyphs. “The best way to decipher the hieroglyphs is to have lots of people working on the problem, so just make them known,” he says. “I think the same logic applies here.”
Douglas Vakoch, the president of METI International, has spent a lot of time studying the problem of news translation. You can tell from his organization’s acronym, which stands for “Messaging Extraterrestrial Intelligence.” He says AI can play a supporting role in detecting and deciphering alien messages, but is not the main role.
“We need to be aware that when we humans try to find patterns hidden in radio statics, we might start with a few simple guidelines that are very similar to the clear rules of AI. But often we don’t see exactly why our rules are inadequate because we don’t spell them out clearly,” Vakoch told me via email. “AI forces us to be aware of how we’re trying to solve problems, and by simply learning from AI how it tries to solve a problem, we can say, ‘You missed something critical. You need to do this instead.’ ”
In his opinion, discovering an alien message is only half the battle.
“An even greater challenge will be to understand what it means. And this is where humans will continue to play a role, even as AI becomes more sophisticated in the coming years,” said Vakoch. “Deciphering a message from aliens will be much more unclear. AI could help us spot patterns in alien messages that humans would miss, but we will still need people to figure out what the message means.”
How long will it take to make contact with aliens? Do we have to wait for a life-threatening mission to a distant star system? More than 20 years ago, Shostak predicted that we would find evidence of aliens by around 2025. And he has been putting a lot of coffee on it for more than 15 years.
Now Shostak admits he may have to pay. “The next time I see you, I’ll buy you a cup of coffee,” he says. “We haven’t found them yet. … Maybe it was just wishful thinking, but honestly I think it was more based on the known rate of improvement in the alien discovery experiments.”
Maybe SETI astronomers just need more time to use Moore’s Law and AI. Perhaps it will be another 20 or 200 years before the promise of Project Hail Mary is put into action and brought into contact with extraterrestrial travelers. But in the meantime I’ll have the cup of coffee.
Here come the spoilers
If you haven’t read Project Hail Mary yet, it may be difficult to keep track of the film’s scientific twists. Some of these plot twists have interesting parallels to real-world science, and I can’t help but point them out.
Project Hail Mary is set to hit theaters on March 20 and is already receiving rave reviews. For more from Seth Shostak, check out Big Picture Science, the podcast he co-hosts. and look for his columns in Astronomy magazine.
My co-host for the Fiction Science podcast is Dominica Phetteplace, an award-winning author, graduate of the Clarion West Writers Workshop and based in San Francisco. To learn more about Phetteplace, visit DominicaPhetteplace.com.
This report was originally published on Cosmic Log, home base of the Fiction Science podcast. Stay tuned for future episodes of Fiction Science on Apple, Spotify, Player.fm, Pocket Casts and Podchaser. Fiction Science is included in FeedSpot’s Top 100 Science Fiction Podcasts. If you enjoy Fiction Science, please rate the podcast and subscribe to receive notifications for future episodes.
