Recently, AI company Synthetaic’s founder and CEO Corey Jaskolski has created a tool called Rapid Automatic Image Categorization (RAIC) that allows users to find almost any large object in satellite photos of Earth within a single day. This could give the National Geospatial-Intelligence Agency (NGA) a great advantage on the battlefield. Jaskolski’s concept is to have RAIC ingest the entire Planet daily photo of the Planet’s surface and process all of it in less than 24 hours.
This type of satellite image scanning was demonstrated by Jaskolski during the Chinese weather balloon incident in January 2020. He applied RAIC to satellite photos from geospatial satellite imaging company Planet and successfully traced the balloon’s origin to China in the matter of days. This was no easy feat as training AI typically needs billions of labeled images, which the team did not have access to. By using technology for generative AI transformers and Generative Neural Systems (GaNS), Jaskolski was able to teach the tool to look for the balloon using a single hand-made drawing.
NGA has taken control of Project Maven, the Pentagon’s flagship AI program for image analysis and is experimenting with geographical intelligence AI programs that employ large language models. Microsoft President Brad Smith believes that the next generation of AI, which would employ GPT-based technology, will allow people to query the Earth.
Synthetaic is an AI company that seeks to revolutionize intelligence gathering and shape the future of technology through the development of generative AI tools. The company’s mission is to solve one of today’s biggest challenges – to make information accessible, actionable, and valuable through automated intelligence for teams that scale. With offices in San Francisco, Seattle, and London, the company offers consulting and education in AI, GANs, ML, and robotics. At Synthetaic, the research team focuses on developing generative AI technologies to enable large-scale analysis of satellite imagery and other data.
