Renee’s four golden rules of artificial intelligence

So much hype, so much uncertainty, so much information about artificial intelligence. Here are some guiding principles… 

Years ago, I wrote a handout, which then turned into a blog post, entitled Journalism ‘legals’: Renee’s golden rules.

My audience was online subs (sub-editors), people who may or may not have had court reporting, or any reporting at all, in their backgrounds. My thinking was to distil what I had learned as a court reporter into a set of easily understandable guidelines that would help people steer away from trouble.

Trouble in that case consisted of contravening either laws or generally accepted practice for the way in which a report about a court case, or an incident involving the law, was presented in a news publication.

The practice of reporting on the courts is generally fairly well codified in any given country, and reflects whatever the legal system of that country is – so South African writers and editors need to know about defamation, whereas their counterparts in the UK need to understand libel.

My post today wades into much less codified waters – best practice when it comes to using AI tools like ChatGPT. I’ve been using ChatGPT and various other AI tools for a while now, and have developed a set of rules that I follow, based mostly on common sense, but also on years of journalism experience.

My golden rules for using artificial intelligence are these:

Stay in your lane

AI is not a human being. It will always try to answer your question. If it doesn’t know the answer, it may make something up. All this is well-known and is often used as a reason not to use AI, ever. But here’s the thing: if you use an AI tool to try to build a rocket ship, you are only going to be able to tell if its output is useful if you yourself know something about building rocket ships. So the rule is this: only use AI tools to amplify your work in your own areas of competence. It’s simple really: your human knowledge and AI together can get lots done. But your lack of knowledge and AI’s hallucinations are a recipe for disaster.

Keep your critical mind alive

Local publication GroundUp reported on January 8, 2025 that a South African law firm is facing the possibility of a Legal Practice Council investigation for allegedly using “Google” and artificial intelligence (AI) to source non-existent legal citations in court proceedings.

The report on GroundUp is detailed and worth reading in full but this comment from the judge is the foundation of my golden rule number 2: “[the lawyer] blindly relied on authorities provided to her [by the clerk] without checking the references.”

It’s simple but headache-inducing: if there’s a reference or a fact or a quote or a time or a place mentioned in text provided by AI, you’re going to have to check it against your own knowledge or against other sources.

Don’t expect AI to do everything for you

Following from the above, it’s clear that putting all your school or work tasks into AI and using the answers without employing your own brain is not a sensible thing to do. The smarter thing to do is to think through what you need done and then ask AI to do the things it does well. An example – do your research, write that report and then ask ChatGPT to check the spelling and grammar for you. Or break down your project into its constituent parts and put that list into AI and ask for help in putting together a week-by-week task breakdown.

Treat AI tools as if they were social media

It’s tempting to use AI to make your job easier. But if you upload information to an AI tool, the question of privacy and security is murky. ChatGPT’s help centre article on how data is used says this: “When you use our services for individuals such as ChatGPT, DALL•E, or Sora, we may use your content to train our models.” (The same page has instructions and links on how to opt out. And scroll down in this article to find out what a “model” is when it is at home.)

Broadly, that sentence from the help article means the content you have uploaded will go into the vast pool of data that Open AI uses to train its tools. It may not matter much if your primary school child’s history homework is “out there” but it might matter very much if your company’s plans for the next year are uploaded and then turn up in the answer to a query by a competitor..

In essence, the extent to which our information in the world of artificial intelligence is private is unclear. Don’t feed anything into the machine that is not public information, or that you would mind your boss knowing about. Just don’t.

THE QUESTION YOU WERE AFRAID TO ASK FOR FEAR OF LOOKING SILLY

What’s a “model” in AI training, Renee?

 Applied research and innovation firm The Decision Lab has a good overview of the concept of an AI model, which they define as: 

Artificial intelligence models are tools and algorithms designed to train computers to process and analyze data in a manner similar to human cognition. These models enable machines to learn from data, recognize patterns, and make decisions with minimal human intervention. 

This example they give is helpful: 

Google Maps and other navigation applications use artificial intelligence models to guide us to our destinations. The machine remembers the edges of buildings that it learned by using data from other travellers and through inputted data via an algorithm. As people use the application on a day-to-day basis, the model incorporates the data gathered from these travels and can give more accurate route information by recognizing changes in traffic flow.

I recommend the rest of the article for a dive into the field.

OTHER THINGS I’VE WRITTEN

Journalism ‘legals’: Renee’s golden rules

Why cultural competence is a thing when using AI tools | Safe Hands

Editing tips and tricks: Take nothing for granted | Safe Hands

Main picture: Nahrizul Kadri, Unsplash

How can I help you make order from chaos? 

Contact me via email.

To get articles like this one in your email, sign up for my Sensible Woman’s Guide to AI and Content Creation, which comes out fortnightly.

Or sign up for my personal fortnightly email newsletter In Safe Hands (it’s a place to pause and think about things).

Book a free half hour of my time here. Find out how I can help with writing, editing, project management, general clear thinking, technical hand holding, or an introduction to AI.

Comments are closed.