AI, the most spoken of technology of the modern era. This is how I define it. And the particular tech confident it operates with so much confidence and sometimes while I'm using it, I get amazed how it can generate and answer to a question that is so wrong.
It makes hallucinations look so factual I sometimes laugh. You could ask chatgpt a about a fruit or anything else that doesn't even exist and it can give you properties and details about the thing as if it actually exists.
This level of confidence in giving the wrong answer is not as a result of some codes of pride written as part of the AI program. The reality is that, chatgpt and all other models are not exactly aware of what they're doing.
Just because it can give you facts doesn't mean it knows just like you do that what it's giving you is factual. It has merely been designed to combine words and try to make meaning to you. AI can does not necessarily know that it's confused about a matter it just keeps going. This is why I think Human involvement will always be needed regardless of the level of advancement we achieve with AI.
We need to keep our hands on the steering wheel no matter what the technology is capable of doing. Perhaps by the end of this year, openAI and other AI companies would have achieved great things but I'm not getting my hopes up and starting a new habit of fully relying of what an LLM or an AI chatbots tells me without doing my own research right after.
Sourcing the image appropriately
https://pixabay.com/photos/robot-chatgpt-class-school-board-7720802/