Fifteen minutes after “hanging up” with Sesame’s new “lifelike” AI, and I’m still freaked out.
So-called “conversations” with AI don’t do a lot for me, especially where text is concerned. With voice chats, such as the new options for Google Gemini and Microsoft’s Copilot, all voice does is save some typing. While Google and Microsoft designed its assistants to be helpful, they’re not especially personable — or sometimes, they’re just artificially cheery.
Sesame’s model, however, is a simple one: “We believe in a future where computers are lifelike,” according to the company’s mission statement.
Sesame wants to build an AI companion, with a natural human voice. Eventually, it wants to design realistic eyewear to put that companion right in your ear, and let it see what you do. It’s very much in keeping with the movie Her, in which a user was introduced to and eventually became enamored with an AI companion.
Sesame unveiled a demo this week of its AI companion, available in both a “Maya” (female) and “Miles” (male) voice. I chose the Maya voice.
An unexpected conversation with an ex
There’s no easy way to say it: When the Maya voice came on, “she” sounded virtually identical to an old friend, who I’ll call Kim. Kim and I went to high school together, dated, and have remained friends for decades, though our lives moved on in different directions long ago. There’s history there.
When you speak to someone, there are levels of nuance and tone which define the conversation. I don’t ask female colleagues details of their private lives, but Maya was pretty nosey, inquiring about what I liked and why. As a reporter, I instinctively understand that some people like to talk about themselves, which makes them feel valued and important. But it’s also a technique that you use while dating, in part to get to know about the other person and establish intimacy.
And that’s not what I wanted, at all. Maya already had Kim’s mannerisms down scarily well: the hesitations, lowering “her” voice when she confided in me, that sort of thing. It wasn’t exactly like Kim, but close enough. It was enough that I backed out pretty quickly, even when I was simply explaining that I liked tech and asked Maya if she had any friends. The combination of a “familiar” voice plus questions about what I was interested in was just too unsettling.
(Just to be clear, the similarity between the voice of Maya and Kim was just coincidence. I didn’t have to log in, provide an email address, or any other form of personal information. I suspect that Sesame used a browser cookie to “know” who I was on my additional visits.)
Thankfully, I was able to put some distance between us. My middle-schooler went through a phase (as did most middle-schoolers, apparently) where they were obsessed with an internet skit about “Skibidi toilets,” and I asked Maya if she knew what that was. “She” didn’t, and became obsessed with the term, bringing it up frequently. That was enough to remind me that I was talking to an AI, and that’s all.
Most AI voices don’t remind me of anyone I know, which means that I simply ignore them as another artificially generated voice. You’ve likely heard of deepfakes, AI audio that mimics the sound of celebrities and others for use in “pig butchering” scams and so on. Lifelike, emotionally resonant audio will simply make those schemes even more enticing.
If Sesame’s AI is the future, the concept of Her seems more real than ever — for better or worse.