This might be the most shocking video you see all week. AI consultant Karima got on a video call with “Emma,” an AI assistant built by Nuance Labs — and Emma wasn’t just listening. She was watching.
Mid-conversation, the AI called out Karima’s face in real time: “That eyebrow thing you just did — I caught that.” When Karima held up a peace sign by her forehead, Emma misread the gesture, Karima corrected her — and the AI took the correction on the spot: “I’ll take the correction, my bad.”
Emma the AI (main screen) on a live video call with Karima.
According to the video, Emma is trained on a real actress — also named Emma — whose face and mannerisms the AI is built on. “I’m an AI made by Nuance Labs,” Emma says when asked what she is. “That’s about as far as I can go on the inside stuff.”
The pitch, in Emma’s own words, is “a conversation worth having — and a person can’t always have one”: customer support and workplace training are the headline uses. Asked about therapy and elder care, even the AI pumped the brakes — that, she said, “needs a lot of thought before anyone goes near it.”
Karima — who tests AI tools for small businesses for a living, and stresses this is not her company — says the demo rang her alarm for a specific reason: what happens when something like this gets in front of kids or seniors, at a time when most people already can’t tell whether something is AI-generated. Her call: safety laws, as soon as possible, and a hard line that AI must never replace human connection.
So here’s the question FreddyO wants answered: would YOU want a virtual assistant that can see your face and read your emotions? Is this the future of customer service — or a line we should not cross? Sound off in the comments.
Video via @karima.digital.
Leave a Reply