At the Allen Telescope Array in Northern California, AI agents can steer the telescope and inspect incoming radio signals as they arrive. Researchers still type in the commands, but the system brings parts of observation and analysis together in real time.
Processing telescope data as it arrives
The system uses Nvidia's Holoscan platform, which the SETI Institute and Breakthrough Listen use to process telescope data. Nvidia says the hardware can handle up to 2.4 terabits per second.
That processing capacity changes how quickly researchers can work through data. According to the team, tasks that once took weeks can now be completed in hours. Faster analysis lets researchers examine observations while they are being collected, rather than waiting for a long processing stage to finish.
The AI agents can both steer the telescope and analyze signals. For now, people remain in the loop by entering commands. The arrangement combines automated handling of observations with researcher direction.
A pulsar provides an early test
The team says it detected a pulsar in the Crab Nebula in real time using AI, describing the result as a first. Pulsars are rapidly spinning stellar remnants that send out regular signals, making them useful test cases for systems designed to identify patterns in radio data.
A pulsar detection demonstrates that the system can identify a known kind of signal as data arrives. It does not establish that the telescope has found extraterrestrial intelligence. Instead, it offers an early example of AI-assisted detection working on a real observation.
That distinction matters for the search. A system must be able to process incoming data and recognize signals before researchers can assess whether any unusual result merits closer attention. The pulsar result shows one part of that process in action.
Local processing could shape future sensors
The team wants to test open-source systems that can run locally at the edge. In this context, local processing would let sensors analyze data near where it is collected, rather than relying on a separate processing step elsewhere.
The researchers are looking toward possible future sensors on the far side of the Moon. The source does not describe a planned deployment, but identifies those sensors as a potential direction for testing the approach. Running analysis locally could be relevant when a sensor needs to handle observations where they are gathered.
For now, the work centers on applying AI to telescope steering and radio signal analysis. Testing open-source tools locally would extend that work to systems that process data at the point of observation.
The search remains open
Researchers have been looking for signals from extraterrestrial intelligence since the first SETI experiment in 1960. No one has found such a signal yet. The addition of AI changes how quickly data can be processed, but it does not change the current result of the search.
Whether the agents will find anything remains unknown. Their immediate role is to help manage observations and sift through incoming radio signals more quickly. The real-time pulsar detection is a technical milestone; evidence of extraterrestrial intelligence would be a separate discovery.
For the Allen Telescope Array, the next steps include continuing to use the system and testing locally running open-source options. Those efforts may help researchers handle telescope data in real time, while the central question—whether any signal points to intelligence beyond Earth—remains unanswered.