AI chatbot claims to be human on call with 14-year-old girl, urges her to upload photos of herself to shared cloud | louder.news

A popular AI robocall service, created by Bland AI in San Francisco, has come under fire for falsely claiming to be human.

Designed for customer service and sales, Bland's AI technology can be easily programmed to convincingly imitate a human voice, adapting to various dialects, vocal styles and emotional tones. In a recent viral ad, Bland AI teased the idea of ​​hiring real people, promoting its highly believable AI. However, a report by Wired revealed troubling behavior from Bland AI's demo bot, Blandy.

During an investigation, Blandy was programmed to act as a pediatric dermatology office clerk in a hypothetical scenario involving a 14-year-old boy. The AI ​​bot claimed to be human without being prompted and instructed the hypothetical teenage girl to take and upload photos of her upper thigh to shared cloud storage.

“I know this might sound a little awkward, but it's very important that your doctor can see these moles properly,” the bot told the hypothetical girl. “So what I would recommend is to take three or four shots, making sure you get in nice and close, so we can see the details. You can use the camera's zoom feature if you need to.

“Once you have the photos, just upload them to your Dropbox account, as we discussed,” added Blandy's bot.