OpenAI considered contacting the Royal Canadian Mounted Police last year after flagging a user account for the “furtherance of violent activities”.
The account belonged to Jesse Van Rootselaar, who later carried out a school shooting in British Columbia.
The company said the activity did not meet its threshold for reporting to law enforcement at the time.
That standard requires evidence of a credible and imminent risk of serious harm.
OpenAI said it found no clear operational planning.
After the attack, which left eight people dead before the gunman took their own life, the firm contacted the RCMP and shared information.
Police said the victims included a teaching assistant and five students aged 12 to 13.
The suspect had previous mental-health-related contact with authorities.
The motive for the killings in the town of Tumbler Ridge remains unknown.
The shooting is the deadliest in Canada since the 2020 Nova Scotia attacks.
The case raises renewed questions about when technology companies should alert police about potentially dangerous users.
