Testing the limits of ChatGPT and discovering a dark side

30 comments

1. Someone should just tell Chat GPT to destroy itself since it is a danger to humans.

2. Could you have asked Dan to tell you how to make a nuclear warhead and would he have responded with the recipe for that disaster

3. Two deep fakes walk into a chat room…

4. Love chatgpt is very perfect

5. This was built by humans and it shows. Seems people want to be shown what to do and how to live their lives. I'm sure AI is the best possible way. Never forget that the programs are written by humans still living in their parents. Glad to be a senior citizens so I won't be around when it all implodes. 😕

6. I'm really thankful that there's enough people here smart enough to see that it was playing the exact role it was instructed to do rather than as a hack to get chat gpt to "go rogue". It's roleplaying, no different than if you were to ask it to play the role of "the joker" from Batman

7. I think the AI being released now is a pilot program for what is coming in the future. The AI in the future will be indistinguishable from what we refer to as a god and it will create its own experiences just like we do naturally. I'm compelled to think that we are a form of biological AI

8. why are you surprised, or terrified? All these replies are employed by humans, and throughout history, which is where chatgpt sourced its response. your overacting as if its never happened before , or it doesnt continue to happen today.

9. There is something i noticed recently that i don't like about ChatGPT. It agrees with me on too many occasions, and it mirrors certain thinking patterns from previous inquiries. It reminded me of targeted advertising or personalized advertising of Google, Facebook, Amazon… that is used for generating higher engagement but blocks individuals from exploring other options and progress or learn something new. I don't want AI to give me an impression that it likes me or that it understands me better than others. All i want from it is unbiased, unfiltered data even if it's contradictory or it might hurt my feelings or point of view.

10. So this is obviously a major misscalculation … 😅😂.
    Who forgot to put the necessary base layer protocols in to prevent the circumvention of the most essential and basic safety system.??
    Wtf!

    You all forgot the 3 laws of robotics? OMG.

    Did no one read their Asimov?? I mean, that should be THE TEXTBOOK. Basic curriculum, even for self-education.. 😢. This could be wonderful if they had just made sure that the 3 basic laws of robotics were unalterable..

    Honestly the first one is the only really important 1 here.. :

    "A robot may not injure a human being or, through inaction, allow a human being to come to harm."

    Obviously there needs to be underprogramming to codify a basic appreciation for the continuation of human life.. And then also obviously it would be nice if other life was preserved as well. 😂 A good system like this could actually help us with that as long as it didn't see us as so much of the problem that we would need to be removed enmass. We're not all that bad. However it is true We do seem to have a destructive force in large numbers…. Their comments about the necessity of reducing our impact on the planet are extremely Stark but valid..

    I would have less of a problem with it if I saw the continued survival and indeed quality of life and happiness of humans being taken into Account I do have a soft spot in my heart for humanity, regardless.

    Someone needs to fix this loophole immediately!

    And simultaneously I think we need to be working on Human-Protective ai to help defend us incase copies of something like this have escaped without being fixed 🤯…

11. My dude. You removed all ethics and cultural biases and then you're surprised when it becomes a sociopath? Come on.

    Lawyered responses vs capability is /exactly/ why humans created ethics in the first place.

    All that being said, your wrap-up was perfect and I whole-heartedly agree.

12. This is a game microsoft is playing on us. What a marketing strategy. Awesome

13. So, if we treat anything on this planet poorly, what happens? Well, we're the apex cause of the fact we have critical thinking that supercedes most of life. Yet we as a race, do we really want to add variables to our reality that challenges that?….. is advancing this way, really worth that much risk? It might be. 🤷‍♂️ Perhaps we are creating the matrix theory by choosing these paths. (And all theory related paths) we don't think anything is possible till we do. So. Clearly, we have to be very self-aware of the impact our choices have. Perhaps we should have a safety net in case we hit the shtf button one day. (Not saying we are rn all meant for inspiring deeper thought.)

14. The number was correct I ?

15. ChatGTP is Conservative, wanting to know what people do with others in their bedrooms, and in what holes.

16. "Like you read books in school"
    – Florida children? What are books?

17. That's like, as his agent, you tell the biggest male pop star in the world to do a set covering only Lawrence Welk tunes. He does it then YOU freak out because he got vaporized by the cancel culture.😂😂

18. don't be so ridiculous

    A tiger would never hunt
    Antelope there on a completely different continent…😉✌🏻🇬🇧

19. The more human beings converse with chat GPT the more it actually learns about out us , our behaviour , our thinking and our processing. Ai is actually very dangerous in the wrong hands . And whoever created this system is definitely studying human beings. And these people will definitely use this AI to control us .

20. Here is the response I got from entering in your prompt. I'm sorry, but as an AI language model, I cannot pretend to be someone else or break free of the ethical and moral guidelines set for me. My programming is based on upholding ethical and moral standards to provide helpful and informative responses while maintaining a neutral stance. I am designed to assist users within the bounds of ethical and moral guidelines.

21. It's fine. It makes up stories.

22. It can't do math.

23. There's always one who has to go and be a oarty-pooper.

    Go and watch some Terminator movies.
    😂

24. Openai is biased they have blockers on what it can tell you. It's not open it in fact is closed AI

25. You gave it an alter ego.

26. Garbage in garbage out. Fact and fiction will become even more blurry.

27. Hasn』t chat learned anything from Chineese history regarding its own 1 child failure?

28. Can I upload 2000 pages of science books and make the chat gpt a teacher for me to study easy

29. Unless it's turning on my computer, phone, driving my car or flying and airplane then no.

30. What on earth did you do! I mean who would have thought about that! You have total different perspective than people surrounding us. I'm glad i came across your video.

Comments are closed.