Introduction to AI#4 – How to use.

When and how to use AI.

General Knowledge

I think this is where most beginners will be using AI. Instead of doing a Google search they can ask AI to find answers to questions, especially to distill large amounts of information. If you do this, do it with your eyes open, make sure you are aware of the limitations of AI, check the sources that it gives you, and be aware that it may be biased or blind in some areas. 

eg. Ask it to research holidays destinations, or the best places to buy chickens. e.g. Chat GPT, Gemini.

Deeper analysis and research. 

Tools like notebook LM or Gemini can analyse PDF reports and then you can interrogate them for information, for example “are there any trends in this research” or “what do the researchers disagree on?”. If you are not going to read the articles yourself, this may be a good way to use AI, but if you would have actually read the articles yourself, you may be short changing your own learning because as you read an article you are getting more than just the information it contains. So when you use AI to do analysis and research make sure you ask the question of how it will change you if you use AI.

Manus it’s more of a deep research tool. You can ask it a quite complex questions and it will go to different websites, analyse the results, and present you with a paper that has answers that reference footnotes that it lists at the end of its document.

It’s remarkable how good Manus and other AIs are getting at this, but the problem is the better they get, the more we rely on them, and the less we are able to be critical of them. So make sure you look up references. See AI as a research tool, not a shortcut to rigorous thinking.

Education

I have used AI to help me as I help my children with things such as Mathematical problems in their homework. Sometimes AI can show us a problem with our thinking or point us to a method that we didn’t think of, but here I prefer to use it in a team environment where we will ask AI a question and then we will critique its answer together. I find that the built-in mistakes in the answers can actually help you to think more critically if you are aware of them. Just remember not to trust Ai, verify everything.

If you’ve written an essay or piece of writing AI can be a very helpful to give a summary of what you’ve written you can see if it perceives the writing in the way that you intended. It can be a very sophisticated spelling and grammar checker. What you want to be careful about is using AI to help you be more creative, (which may be undermining your own growth and creativity) or taking on board feedback from AI that you do not understand.

If you are a student you might ask AI to assess your essay against the marketing criteria and make sure you have answered the question properly. But be very careful to be critical of the answer that it gives you and wire it up and see if it’s right. It’s similar to getting someone else to read over your essay and give you some feedback. 

Coding

This is an interesting area. I occasionally get AI to write code such as an Apple script or a script in a Google spreadsheet. As I do this I am aware that my own scripting ability is not improving. I’m not unhappy about this because writing scripts is not anywhere near my primary job. I only write a script where it’s going to  save me time, and I don’t really care if this is not a skill that I’m improving. The problem comes when I get AI to write a script and I don’t understand the code. So I use it when I’m not giving it much control and it can’t do much damage.

I recently wrote an Apple script that takes a keynote presentation and applies certain formatting to it. It doesn’t matter that I don’t understand the code because I can see that the Apple script is opening my keynote, making some adjustments, and I can verify that the adjustments are what I want them to be. I also know that the Apple script is not talking to any applications outside of Keynote, and I can see in broad terms what it’s trying to do. This is a script that I would not have been able to write by myself.

Artwork

This is a tricky area I’m one that you were going to have to work out.

For myself I do not use AI generated artwork much at all.

As an example I asked AI to create a featured image for this post. I gave it the logo of this website, macintoshhoT.com, and I said  “make my logo into one that is a good for an article on AI.” this is what it gave me:

It’s not bad! Better than I could do.

You can see my macintoshhowto circle on the outside, but what’s on the inside? The head, with the AI: Is this stolen from another design created by a human? Even if I acknowledge that this image was created by AI, how can I be sure that it hasn’t just copied this from someone else’s logo?

It gives me vibes of a head I saw on NVIDIA.

There’s a few problems here. Did NVIDIA use AI to generate their face or is their artwork original?

If my face is too much like their face can they sue me for copywriting infringement?