An elderly Scottish woman was left shell-shocked after Apple's AI-powered voice to text transcription service turned a routine voicemail into a profanity-laced message.
Louise Littlejohn, 66, from Dunfermline, received a call from Lookers Land Rover in Motherwell inviting her to an event. However, when the voicemail was transcribed by Apple's AI tool, the AI included profanities and coarse language, including referring her as a "piece of s**t" and questioning her sex life.
The distorted message on her iphone read
"Just be told to see if you have received an invite on your car if you've been able to have sex. Keep trouble with yourself that'd be interesting you piece of s**t give me a call."
Initially believing it was a scam, Ms. Littlejohn recognized the caller’s area code and recalled that she had previously purchased a car from the dealership. In an interview with BBC, she shared that while she felt that the message was inappropriate, she overall found it quite funny.
AI Struggles With Accents and Context
Experts suggest the mishap likely stemmed from Apple's AI struggling to accurately process the worker’s thick Scottish accent or background noise at the dealership.
This incident adds to Apple's recent AI challenges. Just weeks ago, the company faced criticism after its speech-to-text tool mistakenly transcribed "racist" as "Donald Trump." In January, Apple suspended its AI-generated news summaries after they began displaying misleading notifications.
Apple isn’t the only tech giant grappling with AI missteps. In November 2024, Google’s AI chatbot, Gemini, made headlines after it allegedly told a Michigan student to "please die" while assisting with homework.
The chatbot’s response reportedly included statements such as:
"You are not special, you are not important, and you are not needed. You are a burden on society."
These incidents highlight the ongoing challenges of AI in accurately understanding human speech and intent, raising concerns over its reliability in everyday applications.