We can look to the stars to see how NASA is currently implementing generative artificial intelligence (AI) in its world-class aerospace technologies.
With more than 100 million users, ChatGPT may be one of the highest-profile examples of recent AI technology – but it is just one of many in a larger generative AI movement. Generative AI applies Natural Language Processing (NLP) to Large Language Models (LLMs) so people can use simple prompts to create word, image, video, and even music responses from an infinite repository of data.
And there’s a major component of generative AI that gets lost in translation: processing huge amounts of data quickly and without coding. The applications for generative AI in the aerospace industry are immense; as SpaceNews predicts, “We will likely see more impact from these new models in the next 1-3 years than in the last 10 combined.”
Advanced aviation analytics turn undervalued datasets into gold
Big-name, big-money aviation and defense organizations have already invested in large-scale data collection methodologies, but many have been unable to realize its full benefits due to the sheer volume of data. This applies both to proprietary data collected by national defense and private organizations, as well as massive repositories of publicly available datasets from drones, satellites, and airplanes.
Generative AI is the solution to turning this tangled, overgrown wasteland of data into valuable data for use by data scientists, private organizations, or even the general public. NASA has partnered with IBM Research to use generative AI to create a geospatial foundation model that uses satellite data to enable geospatial analysis three to four times faster than traditional methods.
Generative AI makes hardware and software faster and cheaper
Generative design algorithms take aerospace design requirements — from relatively straightforward geometric and budget constraints to multiple complicated integrated systems — and quickly create design options for aerospace engineers to implement. This supercharged design process can bring new products and systems to market significantly faster than ever before.
NASA’s generative AI-designed aerospace hardware, Evolved Structures, reduces months of work that would take an entire team of CAD designers mere hours with generative AI and the right prompt. The Evolved Structures reduced typical hardware weight by one-third with no sacrifice in performance, and are already being added to space telescopes, Earth-atmosphere scanners, balloon observatories, and planetary instruments.
But it’s not just hardware: Generative AI empowers aerospace engineers with the potential to quickly create C or C++ code for embedded software that will save time and resources. Text-to-code generation will be a major focus for NASA’s applications of generative AI in the aerospace industry and beyond. In fact, it is predicted that machines will write most of their own code by 2040.
And one quantum leap for AI: interactive space exploration
When it comes to generative AI in the aerospace industry, it doesn’t take much to envision Captain Kirk giving instructions to Andromeda or Hal from 2001: A Space Odyssey. And that’s actually where science is about to take us, according to Dr. Larissa Suzuki, who is working with NASA to develop an internal natural language interface for scientists and mission control to directly interact with spaceships and rovers.
The idea is that each system will include a user-friendly natural language interface enabling astronauts and mission control to communicate with it easily. This eliminates the need to search through complicated technical manuals for information needed in real-time. Suzuki envisions astronauts using this interface to obtain guidance on space experiments and intricate maneuvers, making information retrieval and communication more efficient during missions.
“The idea is to get to a point where we have conversational interactions with space vehicles and they [are] also talking back to us on alerts, interesting findings they see in the solar system and beyond,” Suzuki explains. “It’s really not like science fiction anymore.”
Read more:
● Chips Will Fuel U.S. DoD AI Ambitions – And Complicate Them
● Electronic Warfare is Changing — Here’s Where It’s Headed
● Use These 3 Strategies To Prepare For The Next Shortage Market