Insight into the entrepreneur’s AI method of tracking the Chinese balloon and how it could keep spies out of US airspace

By combining satellite imagery, artificial intelligence and a little creative genius, an amazing entrepreneur was able to track the Chinese spy balloon in minutes.

Corey Jaskolski found a way to use his Rapid Automatic Image Categorization (RAIC) software to map the spy balloon’s path from Canada to South Carolina.

4

4

The entrepreneur is the founder and president of Synthetaic, an image and video recognition startup aiming to accelerate the world’s transition to practical AI.

Jaskolski quickly became interested in the recent news of spy balloons over the US and the systems in place that could be used to locate them.

He happened to have an idea that could provide more information about the mysterious objects.

Accordingly WIREDthe high-level engineering process began with something very simple: drawing.

Jaskolski pulled out his drawing material and sketched what he thought the surveillance balloon shot down by the US might have looked like from space.

He then fed his sketch into algorithms developed by Synthetaic. In about two minutes, the algorithms revealed the locations of the balloons.

“I couldn’t believe it,” Jaskolski said.

That was just the starting point. From there, the engineer collected more data from social media sightings and wind models to feed into his software.

“We drew a wide arc through time and space and started the search,” added Jaskolski.

And over time, he and his colleagues were able to create a complete map of the balloon.

“We can draw a 1km trail across the entire United States and just follow the balloon,” he explained.

“We have a trail from Canada to South Carolina where it was popped, with six points along that arc.”

Jaskolski’s findings are not only a notable step forward in the world of AI, but could also be crucial for the fields of intelligence and security.

Holland Michel, a fellow at the Carnegie Council, explained that the future of surveillance rests on this type of technology.

“Combining AI with satellite imagery is undoubtedly a very powerful technology for surveillance, espionage and counterintelligence,” he said.

Jaskolski’s method is still new and needs to be perfected.

The RAIC software can still be modified by human error, leading to false positives if the original human drawing was not accurate enough.

Jaskolski is also aware that there is still room for improvement, but he is proud to be part of this project.

“This human-machine collaboration is my vision of how AI works today,” he said.

“And that’s definitely how we build our product.”

The AI ​​expert hopes to continue using the balloon-tracking technique, no matter how “resource-intensive” it becomes.

4

4