
Joshua Tyler built the world’s first AI network that can learn how to encrypt itself
The University of Tennessee at Chattanooga student submitted their invention disclosure to the University of Tennessee Research Foundation (UTRF) for a provisional patent.
Innovation never seems to slow down for Joshua Tyler, a doctoral candidate at the University of Tennessee at Chattanooga (UTC). Over the past year, Tyler has rolled out several new technologies, with his latest earning attention for its game-changing potential across industries.
This time, Tyler has developed what UTC calls the “world’s first usable Artificial Intelligence (AI) network that can learn how to encrypt itself.”
Exactly one year ago, teknovation.biz interviewed Tyler to discuss a different innovation. At the time, he had just finished his participation in the Innovation Corps (I-Corps) hosted at the University of Tennessee, Knoxville (UTK). He went through that program hoping to commercialize a code he developed for higher-quality audio.
Now, he has taken innovation one step further by creating AI that can to secure its own data.
“It took months of trial and error with traditional architectures, trying different training scenarios and metrics,” Tyler said. “But when it finally worked, it was surprising. You go through months of failure, and then suddenly, something new you invented actually works. That’s when the real fun begins.”
The implications of Tyler’s discovery extend across industries—including defense, health care and finance.
“In the military, they want to do a lot of unmanned vehicles, and a really big consideration is communications,” he said.
New aircraft are being designed to be unmanned because it costs a lot of money to put a human pilot in the cockpit, and every time a human flies it’s a life-risk. However, if these large cars and helicopters are going to be flying solo, then the data being transferred to and from them needs to be encrypted. Otherwise, other countries could hack into our vehicles and aircraft.
Tyler described a real-world incident where this happened. An adversary intercepted the data from an unmanned plane, resulting in the loss of the aircraft.
Beyond the military, Tyler’s AI encryption system could protect sensitive health care information. For instance, in hospitals, data about a patient’s vitals could be intercepted if not encrypted. Tyler’s technology aims to prevent such breaches.
A draft of Tyler’s work has already been published on arXiv, and he is now preparing to submit it for peer-reviewed publication.
At this point, Tyler is considering licensing the technology, finding interested partners, or securing additional funding to commercialize it.
With applications spanning multiple industries, Tyler’s AI encryption network could alter how data is protected in the future, marking another significant milestone in his already impressive career.
Like what you've read?
Forward to a friend!