Joshua Tyler
In a world where data security is an ever-growing concern, Joshua Tyler—a computational engineering doctoral candidate and electrical engineering research associate at the University of Tennessee at Chattanooga—has broken new ground.
Mr. Tyler, who is on track to receive his third UTC degree in May, has developed the world’s first usable Artificial Intelligence network that can learn how to encrypt itself. This AI network, he said, can provide nearly unbreakable cryptography—significantly improving the security of communications.
Mr. Tyler and his faculty mentor, Dr. Don Reising, a Guerry and UC Foundation associate professor of electrical engineering, have uploaded a draft of their publication to arXiv—an open-access repository for scholarly papers. They have already submitted their invention disclosure to the University of Tennessee Research Foundation for a provisional patent.
Mr. Tyler’s AI network, he explained, learns to encrypt data “by transforming an encryption key onto the original unencrypted data. The goal is to ensure that the encrypted message is unique to the key while the original message is still recoverable on the other end.”
“This ensures that when deployed, each encryption is unique and significantly extends the network’s lifespan,” said Mr. Tyler, who received a bachelor’s degree in electrical engineering in 2020 and a master’s degree in 2022.
The research builds on a concept initially proposed by Google, known as Adversarial Neural Cryptography. While Google demonstrated the potential for AI-driven cryptography, its approach faced significant limitations—particularly in ensuring the encryption key’s influence on the encrypted message and additional communication overhead.
“I copied over Google’s setup and trained their network on my side,” he said. “The network was encrypting the information, but we found out that there wasn’t a lot of uniqueness on the encrypted side when we were switching keys, so that makes the overall life of the network shorter. You’d only get to encrypt one message per network.”
Dr. Reising, who has worked with Mr. Tyler for more than six years, said he “basically challenged Josh to go and find a way to get this thing to generate a unique code or a unique encoded message.”
“And that’s what he did,” Dr. Reising said. “He went off and worked on developing his own technique.”
Dr. Reising recalled a pivotal moment during the process.
“I asked him, ‘What architecture are you using? Are you using CNN? Are you using an LSTM? What are you using?’
“And he’s like, ‘No, I’m not using any of those. I made my own.’
“I said, ‘What do you mean you made your own?’
“He said, ‘I made my own and it’s a deep learning network.’ That was crazy and it was pretty awesome.”
By rethinking the structure of AI networks, Mr. Tyler developed a “novel neural network architecture” that addresses these challenges.
The result is a network that offers nearly unbreakable encryption and unparalleled adaptability in safeguarding sensitive data.
“I changed the network architecture so that the influence of the key was still maintained through the entire structure of the network,” he said.
A crucial feature of Mr. Tyler’s system is its rapid adaptability, which allows it to retrain itself in seconds to produce entirely new cryptographic algorithms. This new architecture ensures that each encryption remains unique, effectively overcoming the limitations of previous methods.
“Every time you retrain the network, you get a different cryptographic algorithm,” Mr. Tyler said. “So then, even if you use the same key across two differently trained networks, you’ll get a new encryption scheme.”
“These things train really fast so that we can have a new cryptographic algorithm in about 16 seconds.”