A couple short/intermediate problems I see with not only ChatGPT but “AI” (the definition as used in applications today is pretty broad) in general is (1) built in bias and (2) unintended consequences in human development, or more accurately lack of human development.
Every single human being has biases... we all have a world view shaped by beliefs, experiences, etc. ChatGPT has been created by humans, with not only their language models/algorithms as defined in the code written by these humans, but also content policy intended to be guardrails. Inevitable, those guardrails (which have already been tricked) are going to be determined in part by the creators built in biases. So, this creates, intentional or not, a new path to pass these biases through something that’s “non-human” which in turn presents the opportunity for people to not filter it the same way as they might from another human. Potentially disruptive to say the least, and that doesn’t begin to explore more sinister uses for it.
Secondly, what happens as this tech improves? While you were able to point out some faults with its response, how many others that aren’t as familiar with the subject would be able to? How confident are we that it’s use can be limited to appropriate use cases? What’s to prevent students from employing it to write every single paper they need to? Are we providing the means to short circuit one’s own intellectual development and learning process by providing a tool like this? I don’t know the answer, but I don’t think it’s completely benign in this regard.
The longer term/more theoretical issue is what Rogan and Weinstein discussed. I would argue the true goal of “AI” is to create something smarter than humans. And Weinstein is exactly right... if it’s possible to achieve, how do we know when this intelligence gets to that point? And the bigger question, why should we expect to control it at that point? If you can create something capable of solving problems that humans can’t, it’s also capable of creating problems humans can’t solve. And that’s the real issue, once again it’s human hubris.
Thank you for these additional points to consider and questions to ponder. I think you've raised some that we all should be considering. Now that this tech has been created, only time will tell what humans purpose it for. No doubt that will include positive uses, but also evil as it seems humans have used every invention for sinister reasons eventually. Undoubtedly the inventors of ChatGPT have good intentions in mind, but that is also what the road to hell is paved with...
A couple short/intermediate problems I see with not only ChatGPT but “AI” (the definition as used in applications today is pretty broad) in general is (1) built in bias and (2) unintended consequences in human development, or more accurately lack of human development.
Every single human being has biases... we all have a world view shaped by beliefs, experiences, etc. ChatGPT has been created by humans, with not only their language models/algorithms as defined in the code written by these humans, but also content policy intended to be guardrails. Inevitable, those guardrails (which have already been tricked) are going to be determined in part by the creators built in biases. So, this creates, intentional or not, a new path to pass these biases through something that’s “non-human” which in turn presents the opportunity for people to not filter it the same way as they might from another human. Potentially disruptive to say the least, and that doesn’t begin to explore more sinister uses for it.
Secondly, what happens as this tech improves? While you were able to point out some faults with its response, how many others that aren’t as familiar with the subject would be able to? How confident are we that it’s use can be limited to appropriate use cases? What’s to prevent students from employing it to write every single paper they need to? Are we providing the means to short circuit one’s own intellectual development and learning process by providing a tool like this? I don’t know the answer, but I don’t think it’s completely benign in this regard.
The longer term/more theoretical issue is what Rogan and Weinstein discussed. I would argue the true goal of “AI” is to create something smarter than humans. And Weinstein is exactly right... if it’s possible to achieve, how do we know when this intelligence gets to that point? And the bigger question, why should we expect to control it at that point? If you can create something capable of solving problems that humans can’t, it’s also capable of creating problems humans can’t solve. And that’s the real issue, once again it’s human hubris.
Thank you for these additional points to consider and questions to ponder. I think you've raised some that we all should be considering. Now that this tech has been created, only time will tell what humans purpose it for. No doubt that will include positive uses, but also evil as it seems humans have used every invention for sinister reasons eventually. Undoubtedly the inventors of ChatGPT have good intentions in mind, but that is also what the road to hell is paved with...