00:00A bombshell new lawsuit claims that OpenAI, the parent company of ChatGBT, could have stopped a
00:04mass casualty school shooting. In February, eight people were shot and killed in Tumbler Ridge,
00:09British Columbia. But could app safeguards have made a difference? The suit filed in U.S. federal
00:14court says yes, accusing the tech company of being too greedy to install user guardrails.
00:19The family of one of the victims believes that without those safeguards, ChatGBT was allowed to
00:23advise the shooter leading the transgender individual to take the lives of six children
00:28and two adults. The shooter had so many disturbing conversations with ChatGBT that the account was
00:34suspended months before the attack. After simply setting up a new account, the troubled teen was
00:38back in business, allegedly using the bot to help plan the massacre. The court filings claim that 12
00:44OpenAI employees pushed the company to alert police to Jesse Van Rootzler's disturbing usage of their
00:50product before the attack. But OpenAI did nothing because, as the suit alleges, it would, quote,
00:56set a precedent. OpenAI would be compelled to notify authorities every time its safety team
01:00identified a user planning real-world violence. That point right there, the policing and reporting
01:06of user content is really the crux of this entire issue. The filing goes on to claim that if OpenAI
01:12were to alert authorities to a large number of violent conversations that were being had on its
01:16software, it would require a whole team of people to do so. And as the suit suggests, quote,
01:21the public would finally see what OpenAI was desperately trying to hide, that ChatGBT is
01:26not the safe, essential tool the company sells it as. A company rep told The Post that they have
01:31strengthened safeguards in the app, changing the way it responds to certain prompts and trying to
01:35connect users to mental health services if they seem to be in need. They also said the company is,
01:40quote, strengthening how we assess and escalate potential threats of violence
01:44and improving detection of repeat policy violators.
01:47a
01:47data
01:47in
01:47a
01:47the
01:47the
01:47the
Comments