The military wants to bring the U-2 spy plane into the 21st century with artificial intelligence technology, The Wall Street Journal reported.
According to the news outlet, Air Force reconnaissance experts have enlisted Stanford University engineering and business students to develop advanced computer programs to analyze the old-style Kodak film currently used by U-2s over Afghanistan, Iraq, Korea, Syria, and other hot zones.
If the effort works, computers would be able to quickly scan miles of U-2 film – and count cars, airplanes, motorcycles, buildings, and even individual people, something human analysts now have to painstakingly plow through.
"Our analysts are already saturated with data — too much imagery, too few analysts, and too little time," Air Force Col. Jason Brown, commander of the 480th Intelligence, Surveillance and Reconnaissance Wing, told the Journal. "Imagine if an algorithm could sweep through the data to cue analysts on potential areas of concern. Airmen could spend less time searching and more time making sense of the things they see."
The first U-2 spy plane became operational in 1957 under the aegis of the Central Intelligence Agency; three years later, U-2 pilot Francis Gary Powers was shot down while taking photographs over the Soviet Union, sparking an international incident. In 1962, then United States Ambassador Adlai Stevenson unveiled U-2 images at the United Nations to prove the Soviets had installed missiles in Cuba.
Today, updated U-2 planes that are part of a fleet flown by Air Force pilots conduct combat missions over Afghanistan, Syria, and Iraq, as well as performing humanitarian surveillance – including bringing back images of Hurricane Katrina damage in New Orleans in 2005 and damage caused by the 2010 Haiti earthquake, the Journal noted.
But humans can read only about 5 percent of the images the spy planes capture. The Air Force is hoping computers can learn to scan the images, the Journal reported.
Stanford students working with Air Force experts have already developed a prototype of an automated system to read the U-2 imagery, through a class called Hacking 4 Defense, and the Air Force is expected to test the system soon, the Journal reported.