World · 18 d ago
CC BY-SA 3.0 · Wikimedia Commons
Tumbler Ridge shooting puts AI chatbots' duty to report violence in question
On Feb 10, an 18-year-old killed his mother and half-brother, then six people at a school in Tumbler Ridge, British Columbia, before taking his own life. For months before, he had consulted ChatGPT about the attack, according to reports.
And us?
In British Columbia
In September, the B.C. government sued OpenAI and its CEO, Sam Altman, accusing them of failing to prevent the shooting, and victims' families filed their own suits. OpenAI apologized publicly for not alerting authorities.
On Oct 14, US authorities charged a Washington State man with conspiring with the shooter to plan the attack.
"The chatbot is not a licensed therapist. There is no special relationship between the company and the victims."
— Bryan Choi, law professor, University of Colorado Boulder
What we know
- B.C. and victims' families are suing OpenAI
- OpenAI apologized for not alerting police
Not yet
- Whether AI firms have a legal duty to warn
- Where the line on reporting users should be
Experts say chatbots cannot reliably tell a hypothetical question from a real threat, and that forcing companies to monitor chats raises privacy concerns for millions of users.
