Artificial intelligence development on Michigan State’s campus? It’s more likely than you think.
The National Association of Broadcasters, or NAB, plans to fund it. Michigan State won a grant to start working on an A.I. prototype that could report news in the voices of local broadcasters.
The A.I. in question is DeepTalk, a conversational agent that would be taught to listen to and mimic a human voice.
MSU researchers and WKAR Public Media pitched the concept to PILOT, NAB’s technology innovation initiative, and it won PILOT’s 2019 Innovation Challenge.
Prabu David, dean of MSU's College of Communication Arts and Sciences, said DeepTalk “moves us a step forward,” in the university’s efforts to understand human reactions to computers, chatbots and conversational agents.
“With DeepTalk, a broadcaster can transfer the voice and style of on-air talent to a conversational agent, thus creating compelling human-computer interactions,” David said in a news release.
DeepTalk is a joint endeavor between the College of Communication Arts and Sciences, the Department of Computer Science and Engineering, WKAR Public Media and the Integrated Pattern Recognition and Biometrics Lab at MSU.
PILOT is allocating about $150,000 between MSU and the University of Minnesota, the two winners of the innovation challenge. The University of Minnesota pitched another A.I. concept: Jukebot, a chatbot capable of answering music questions for users and giving feedback to radio stations.
John Clark, PILOT’s executive director, said his organization is looking forward to working with the two universities to develop their prototypes and “provide broadcasters the ability to serve their communities through AI.”