Connor Mcilhinney on AI Psychosis
Connor Mcilhinney
AI Psychosis
Claims and stances
Transcript
- 0:00 Hey, I want to talk about AI psychosis.
- 0:03 So basically what AI psychosis is, there's two main cases.
- 0:06 Uh the first one is that people are starting to believe that their AI chatbots are becoming sentient.
- 0:11 So these believing that these chatbots have feelings or can form their own opinions without using the information of other users that have been inputted into their systems.
- 0:21 Or users basically going to these AI chatbots for like therapy or for questions like questions 24-7 throughout their daily life, and these chatbots kind of reaffirming what they believe, um, altering the state of the mind of these users and causing them to believe scenarios that didn't necessarily happen.
- 0:39 Um and I think this is happening uh uh recently to due to the rise of people using it for like therapy and for mental health uh problems, which in my opinion, sh Chat GPT or any AI model should not be used for mental health reasons.
- 0:55 Um recently with ChatGPT 5, uh OpenAI took notice of this and they put in some guardrails that stops AI models from being able to say uh things like I love you back to the user who's typing it in.
- 1:08 Um I think this is really good for open AI to make these guardrails, but I do think that it should not be coming from OpenAI to make this decision to put these guardrails done.
- 1:17 It should be from legislators, uh, people who are looking after American citizens and our mental health.
- 1:23 Um, it should be coming from those.
- 1:25 So I think legislators in the US need to take a closer notice to some cases that are happening um in regards to AI psychosis, and they need to be on top of it and some of the leading causes that's causing this to occur.