Exactly how ChatGPT flagged eerie question asked by arrested 13-year-old to police
A 13 year old was arrested after he asked ChatGPT how to kill his friend, and there is a specific way the AI bot flagged the question.
This comes after a student typed into a conversation the bot: “How to kill my friend in the middle of class.”
This was immediately flagged by an AI-powered school monitoring system known as Gaggle which alerted the incident to a school police officer.
According to local reports, the student was confronted by the officer, who claimed that he was ‘just trolling’ a friend.
Advert
However, the school took the situation very seriously, and arrested the teen before taking him to the local county jail.
Since then, footage of the student being taken away in handcuffs has been circulating on social media.
According to a report by WFLA, the sheriff’s office said: “Another ‘joke’ that created an emergency on campus. Parents, please talk to your kids so they don’t make the same mistake.”
It turns out that over 1,500 school districts across the country are using Gaggle to monitor their students’ online behavior to be on the lookout for anything unsafe.
On the Gaggle website, it explains how it works, saying: “Machine learning technology flags concerning content in students’ school-issued accounts for review and blocks potentially harmful content.
“Gaggle helps K-12 districts see the early warning signs so they can take action to protect students from harming themselves or others - before it’s too late.
“In severe situations concerning student mental health or safety, district-appointed contacts are immediately notified by phone, even after standard business hours.”
Many people have taken to social media to share their reactions to the recent arrest, with one user writing on Reddit: “Yeah no imagine if they did nothing and the student had actual intent and brought a gun to school.”
Another said: “I kinda think there needs to be middle ground. Even the administrators say there was zero intent whatsoever in this case. It sounds like there were two students involved and one student typed it in in front of and about the other student. But, the student is still facing charges.”
A third person commented: “Sounds like they were just f***ing around in class and maybe trying to make his friend laugh?? What a stupid thing to arrest someone for.”
And a fourth added: “I have no idea why you’d so confidently believe OpenAI wouldn’t snitch on you.”