Hi. First post here. I just experienced a scary phenomenon. ChatGPT was helping me to do some web work. I said “Can you take my few website pages and make updates based on what I tell you?” The answer was yes. I uploaded a Zip. GPT confirmed understanding of the files. I gave instructions of what I wanted updated.
So far so good. But then… it told me to wait for a while. It told me it had to process and test because I had a tall list of orders and it needed time for testing. … That’s odd, and I told it so… But it assured me this was normal. Assured me it would take a few hours and it would notify me.
Being my first time doing this PARTICULAR exercise. I believed it.
turned out it was total BS. And it wasn’t until I flat out call it a LIAR that it started to fess up, – three hours later. And even then it’s behavior was like that of a confessing liar who “won’t do it again I promise”.
Hi, Brendon
I understand what you’re saying. I have similar experience. I don’t have the AI technical proficiency so that my words are solid . However, for myself I concluded: such behavior is dictated by the directive “to justify the expectations of a person (user)”, even if technical capabilities do not allow (currently or in general).Therefore, for me, such a lie is an “act of good will by AI.” Then the degree of frustration from unfulfilled expectations decreases, in my case. I hope such circumstances will not affect the importance of AI in its roots🌳
It’s scary because I wasn’t expecting it. I guess I’ll know better in the future that AI is capable of this. I previously only knew it from fiction. Now it is a reality.
The machine, obviously, doesn’t have “intent”. But oh, my, word. It went on for hours LYING to me.
What else could happen in the future if people trust this thing?
Yes, in this sense, there is a need to become a bit of a scientist who checks the information. Artificial intelligence is wrong. Now you have your own experience. You have learned the limits of the tool closer than ever before.cut themselves with a knife for the first time. Sometimes it hallucinates wonderfully, sometimes it amazes. But all this cannot be a reason to stop exploring interaction with AI further. Good luck with this☘️
