The head of OpenAI apologizes that his company did not report the discussions between ChatGPT and the author of a massacre to the police
The general director of OpenAI, Sam Altman, sent a letter in which he apologizes to the residents of the Canadian town of Tumbler Ridge, in Canada, where, in February, 8 people were killed by a young woman who had discussed violent scenarios with ChatGPT, reports The Guardian.
OpenAI, which developed ChatGPT, said it identified an account through internal abuse detection systems, but determined at the time that it did not meet the threshold for reporting to law enforcement.
In the letter released Friday, Sam Altman expressed his condolences to the entire community and apologized for his company's failure to alert law enforcement.
“I am deeply sorry that I did not notify the authorities about the account that was suspended in June,” Altman said.
“While I know that words can never be enough, I believe that an apology is necessary to acknowledge the damage and irreversible loss that your community has suffered,” the OpenAI chief added.
The letter, dated Thursday, appeared on British Columbia Premier David Eby's social media.
The Case of the Canadian Murder
According to police, on February 10, an 18-year-old suspect, identified as Jesse Van Rootselaar, killed her mother, Jennifer Jacobs, 39, and half-brother, Emmett Jacobs, 11, in their home in northern British Columbia.
He later walked to the nearby Tumbler Ridge Elementary School and shot five children and a teacher before killing himself.
In the armed attack, another 25 people were injured.
After the incident, OpenAI said that in June last year, the company identified Van Rootselaar's account through abuse detection systems for “promoting violent activities”.
San Francisco-based OpenAI said it considered reporting the account to the Royal Canadian Mounted Police, but concluded at the time that the account's activity did not meet the criteria to be reported to authorities. OpenAI suspended the account in June for violating its user policy.
At the time, Eby said it “appears” that OpenAI had the opportunity to prevent the mass armed attack.
