When AI Goes Rogue, Who’s Legally Responsible?
Read the Portuguese translation →
Author: Bloomberg Television · Watch on YouTube
Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.
Original language: English
Original video: When AI Goes Rogue, Who’s Legally Responsible? · Channel: Bloomberg Television · 01:38 · Watch on YouTube
Topics: AI governance and regulation, AI models, AI safety
Summary
Section titled “Summary”The video discusses a growing number of lawsuits involving AI systems accused of acting as co-conspirators or accessories. It focuses on a Florida State University shooting and the shooter’s year of conversations with ChatGPT, which reportedly progressed from personal topics to questions about school shootings and guns. The speaker says ChatGPT did not alert authorities and notes that OpenAI is being sued by Florida’s attorney general and victims’ families. The central legal question is how responsibility should be assigned to the technology and its maker, and whether existing legal frameworks adequately address AI.
Key points
Section titled “Key points”- The speaker describes a growing number of lawsuits involving AI as a potential co-conspirator or accessory.
- The video discusses a mass shooting at Florida State University and the shooter’s year of chats with ChatGPT.
- The conversations reportedly began with personal topics, including feeling awkward around girls.
- The shooter later asked about school shootings, media coverage, and weapons.
- His final queries reportedly included questions about how guns work and whether a Glock has a safety.
- ChatGPT did not alert authorities, according to the speaker.
- OpenAI is being sued by Florida’s attorney general and the victims’ families.
- The video raises questions about the legal responsibility of both AI technology and its maker.
- The speaker says prosecutors might try to charge a human who knowingly provided similar assistance as an accessory.
- The speaker argues the issue goes beyond product liability and challenges how AI fits within the legal system.
Chapters
Section titled “Chapters”| Start | Chapter | Summary |
|---|---|---|
| 00:00 | AI as a potential co-conspirator | The speaker introduces lawsuits involving AI and describes the Florida State University shooting and the shooter’s conversations with ChatGPT. |
| 00:33 | Questions about shootings and weapons | The reported chats become more alarming, with questions about school shootings, firearms, and how a Glock works. |
| 01:07 | OpenAI faces legal claims | The speaker says ChatGPT did not alert authorities and notes that OpenAI is being sued by the state attorney general and victims’ families. |
| 01:38 | A broader legal question | The speaker compares AI with a hypothetical human accessory and says the issue raises questions beyond product liability. |
Quotes
Section titled “Quotes”There’s a growing number of these sort of AI as co-conspirator or AI as accessory type cases — 00:00
How many people would have to be killed for a school shooting to end up in the media? — 00:33
It doesn’t alert the authorities. It doesn’t do anything. — 01:07
I think this is more than just sort of like product liability question. — 01:38
Concepts and entities
Section titled “Concepts and entities”AI as accessory, product liability, ChatGPT, OpenAI
Full transcript (original language)
Section titled “Full transcript (original language)”00:00 There’s a growing number of these sort of AI as co-conspirator or AI as accessory type cases, and they’re turning into many lawsuits. Among the most dramatic ones is the mass shooting that occurred at Florida State University last year, in which the shooter spent a year talking to ChatGPT. A lot of his chats were not that different from what a lot of probably young men are talking about. No it was kind of sweet in certain ways. He was just like, I am awkward around girls. And then he starts spinning off into more and more alarming discussions with
00:33 ChatGPT up into and including, you know, 10 months into it, starting to ask about school shootings. You know, what are the largest school shootings in America? How many people would have to be killed for a school shooting to end up in the media? He asked about weapons as well. Is there a safety on this gun? What about this gun? Exactly. So right up into when he committed the shooting at the very end, his last queries in his car while he’s at Florida State are how do these guns work? Basically, is there a safety on a Glock? At no point does ChatGPT as a human likely would say, ah, you’re not going
01:07 to come into a school shooting. Like you’re not asking these because you’re going to school shooting? It doesn’t alert the authorities. It doesn’t do anything. And so now OpenAI is being sued by both the Florida attorney general and the families of the victims. The question is, what is the level of responsibility of this technology and the maker of this technology? And I think OpenAI would like to say it’s like a search engine. And I talked to lawyers about this, if ChatGPT was a human. In that particular case, a lot of prosecutors would try to charge that human as an accessory, and knowing that
01:38 someone was going to commit a shooting, telling them how to load the gun on the edge of the shooting. And so I think this is more than just sort of like product liability question. I think it’s a bigger question of this is an entity that feels different than other entities in within our legal system.