For this assignment, I decided to read the article which looked at the differences between Chat GPT-3.5 and Chat GPT-4. Despite the many improvements that have been made, there are still many flaws inherent to artificial intelligence. In particular, I found the flaw about artificial intelligence hallucination to be quite interesting. Upon reading the article, I was actually quite surprised by these findings of artificial intelligence hallucination. As someone with little academic knowledge on artificial intelligence, I decided to do an additional reading as to why hallucination happens. The article stated that hallucination could happen due to issues with the input data used to train the artificial intelligence or due to improper transformer decoding in language-based artificial intelligence models such as Chat GPT. This conversation led me to have a similar concern about artificial intelligence in the context of human decision-making. For instance, making decisions on behalf of a judge or for hiring purposes. Many people have suggested that this is an excellent way of reducing bias and prejudice within these domains. However, the issue is that these artificial intelligence models are going to trained using previous human outcomes - which have inherently been prejudice. Hence, if the AI is not being trained on entirely impartial data, then it might simply replicate many of the same biases and prejudices that we see today. I think this is an important consideration when utilizing artificial intelligence in this context. Even though the AI might not hallucinate, it may not ultimately be serving the purpose that it was created in order to manifest. https://www.marktechpost.com/2023/04/02/what-is-ai-hallucination-what-goes-wrong-with-ai-chatbots-how-to-spot-a-hallucinating-artificial-intelligence/#:~:text=The%20phenomenon%20known%20as%20artificial,%2Dworld%20input%20(data).