https://www.axios.com/2023/01/10/artificial-intelligence-hype-explosion-generative-ai-chatgpt

An example is ChatGPT that can create human like answers to general questions. This tool is based on 2021 machine learning from reading over 170 Billion words that generates Trillions of probability connections between those words.

The AI also applies a huge set of discovered rules an probabilities about how humans answer questions on any subject.

Given a detailed question, ChatGPT starts writing and then uses its 570GB of learnt writings to predict the next word in the conversation

This output reads like a human wrote it and often does very well on discussing a topic. However, the more detailed the question, the tool can begin to miss key elements out or make errors of fact

There are other AI generative tools coming out every week that can use text, image or sound inputs to make conversations, speech, computer code, images, videos, music and much more

All these tools are not as creative or have the ability for critical thinking or outputting something genuinely new or ‘out the box’ which many people can do.