With AI tools already being used to manipulate voters across the globe via deepfakes, more needs to be done to help people comprehend what this technology is capable of, says Microsoft vice chair and president Brad Smith.

Smith highlighted a recent example of AI being used to deceive voters in New Hampshire.

“The voters in New Hampshire, before the New Hampshire primary, got phone calls. When they answered the phone, there was the voice of Joe Biden — AI-created — telling people not to vote. He did not authorize that; he did not believe in it. That was a deepfake designed to deceive people,” Smith said during a Global Stage panel on AI and elections on the sidelines of the Munich Security Conference last month.

“What we fundamentally need to start with is help people understand the state of what technology can do and then start to define what's appropriate, what is inappropriate, and how do we manage that difference?” Smith went on to say.

Watch the full conversation here: How to protect elections in the age of AI