Introducing AI

I'm new to AI. This is likely old hat to most people, but I have started using Copilot in windows and Gemini on my phone.

I start by asking if the AI knows about the subject at hand. Then I ask the more detailed question.

Let's say for example I want to know about the use of commas in English. I would start by asking if it understands English grammar. When it responds yes, I ask something like "can you describe the different uses of commas in written English". The answer will then be presented.

The obvious problem is hallucination. The AI only regurgitates what it has learned, if it learned from a duff source then the answer it will give will be duff. The old addage: Garbage in garbage out holds true. The main difference is that the AI presents its information assuming that it is true.

A while ago I was told by a a telemarketer that the way to convince someone about something is to present two facts and then a plausible supposition. The plausible supposition being something that was linked to what she was selling. It's how our brains work, so beware.

I was using AI to set up my internet web hosting. The AI started to contradict itself. I was glad that I knew something about what I was doing before I started.

With that in mind you might be able to make a better decision about the plausibility of AI output. Bear in mind that there's an element of sales built in to the AI, and it will only get stronger in the future.

That aside, the AI becomes a pretty good first pass at getting information together. The models are getting better and the training is getting better all the time.

The AI has become a great way to get information together to learn virtually anything.