What happens when your conversations with AI become more real, more persuasive, than the world around you? Click here to subscribe to our channel 👉🏽 https://bbc.in/3VyyriM Stephanie Hegarty follows the story of three people, in three different parts of the world, who have fallen into an extreme spiral of dangerous delusion while using AI. A man in Japan attacked and tried to rape his wife after months of intense conversations with ChatGPT. A young woman in the US, thought she was part of an underground group helping immigrants to escape from the authorities. And a man in Northern Ireland was convinced by Elon Musk’s AI, Grok that a van full of people were coming to kill him. 00:00 Intro 00:55 Grok telling me it was sentient. 01:44 Using ChatGPT for the first time 02:27 My husband attacked me after using ChatGPT 03:49 Grok tells Adam he is being monitored. 04:43 Everything becoming a symbol for Shauna. 06:03 Becoming main characters in an chat delusion. 06:35 A van of people coming to harm Adam. 07.46 ‘Taka’ believes he has a bomb in his bag. 09:04 Shauna takes a walk with ChatGPT in the middle of the night 10:35 ‘Taka’ returns home and a violent incidence is triggered. 11:45 Research is growing into AI delusions. A professor at King’s College London details results. 13:42 the end of peoples AI delusions journey for now. Check out more videos on AI & Technology, here 👉🏽 https://www.youtube.com/playlist?list=PLz_B0PFGIn4e0fyBwDLpcmhlvfBn6H0Qx Find more of the best BBC World Service documentaries you can watch in less than 30 minutes here ➡️ https://www.youtube.com/playlist?list=PLz_B0PFGIn4cI2qSy69-3UkgBWaXGHRrx ---------------- This is the official BBC World Service YouTube channel. If you like what we do, you can also find us here: Instagram 👉🏽 https://www.instagram.com/bbcworldservice Twitter 👉🏽 https://twitter.com/bbcworldservice Facebook 👉🏽 https://facebook.com/bbcworldservice BBC World Service website 👉🏽 https://www.bbc.co.uk/worldserviceradio Thanks for watching and subscribing! #BBCWorldService #WorldService #aitechnology #aichatbot #chatgpt #grok
ADVERTISEMENT
2:18 why did her voice just drop 5 octaves?
AI doesn't need to be sentient to cause real damage. It just needs to be convincing enough to fuel someone's existing vulnerabilities. Safeguards and digital literacy are urgent. 🛡️
I don’t get why people blindly believe everything ChatGPT says. They seem to lack critical thinking skills.
When mentally ill people encounter AI it seems to make things worse.
We already have big tech negatively influencing people via algorithm to rob them of any and every spare second to serve another ad. They hire psychologists and do research on people to optimise the content served trying to get you to stay that little bit longer. They need regulating yesterday.
The problem is we don't teach critical thinking in schools, instead we just kind of hope they pick it up while memorising the English, Maths, Physics, Geography, History, etc that they need to pass exams. We should have active classes dedicated to Debating, Logic, Philosophy, etc.
The problem is with the people. They’re highly delusional to begin with.
No mention of the prompts these people typed in. 😅
Some people are crazy and Chat GPT is eager to support whatever delusions they have
The moment he said, "Computer don't lie." You are doomed to be manipulated.
Voice change at 2:20. AI?
AI has no real concept of what it is saying. Like I talked to Claude yesterday. It said to me it does not have memory of our conversations. And two sentences later it mentioned something I said about Palantir in a previous talk. It contradicts it self. Its replays are only based on your input, the things you say. It has no moral compass, it does not know what it is saying. However, AI is super good in finding and matching patterns. This is why it is dangerous.
Am I the only person that is not interested in talking with a machine? Never really tried these LLM powered chatbots
I don’t use Ai as a friend or therapist or spiritual advisor. Use Ai as a tool to look up facts and always check the facts and don’t believe everything. If you are prone to conspiracy theories or have mental issues Ai can definitely do more harm than good.
What Shauna experienced is Apophenia,( making meaningful connections and finding patterns in random or mundane objects and stimuli) It usually happens on drugs or high spin 'spiritual' states.
This is why everyone needs to read science fiction
I'm quite grounded, tech savy person, but recently I also got into trouble using AI. The amount of sugarcoating it does just to keep you going is staggering. Was smart enough to double check things, spent some time questioning everything and realized that AI isn't really that smart yet. Just a weird prediction machine.
These people seem like they were delusional prior to using AI. The AI chatbots just made it worse.
The problem is not AI, the problem is people is not grounded. Fake news, gold diggers, scammers, whatever... You have to be grounded and smart, and not just take everything as true. This is essential for humanity as a whole to detect and investigate the real facts to form our own opinions and not take bullshit.
People have no basic critical thinking skills.