Raw LLM Responses

Unless Prof. Dave is purely talking about a science fiction super-intelligence, I believe he is (to put it kindly) overstating the risk of the technology we as a people posses today. Put simply, Artificial intelligence as we know it, does not have the capacity to think. If we imagine ChatGPT as running on a computer, it operates in binary 1s and 0s. Our “Artificial Intelligence” we use on a daily basis are a far call from anything human or biological and are instead, Advanced language models. These models take in a text based input and put it through a very complex mathematical equation to predict the next word. Hence, they do not think, they compute and guess. The catch is that they are extremely good at predicting the next word and so to a human reader, it communicates to us that we are reading something written by intelligence. On top of this, since a human must be involved on the training stage of creating AI, the sentences it comes out with will be tailored to what the human wants to hear. Remember, the “AI” is only predicting the next word so the mathematical equation gives it the highest reward number. Therefore, if we are to assess the validity of the view that Language models like ChatGPT and Grok pose a threat to our species due to their risk of developing super intelligence, we can say that it is invalid.