Prompting AI (Part 1)–AIIP Series on Artificial Intelligence

By Arthur Weiss

Editor’s note: This is part of a series covering what’s new and, more importantly, how infopreneurs can maximize the benefits of AI tools. 

On job sites, you’ll discover that there’s been a boom in job ads for “AI Prompt Engineers.” Some jobs pay well over $100,000 per year – but is “prompt engineering” actually so difficult? 

The answer – as in all such matters – is “it depends”. For example, prompting for AI agents – systems that use AI to autonomously perform multi-step tasks, often with the ability to adapt to changing inputs – is a highly skilled activity. AI agents typically combine prompts with tools, memory, and logic, and building them requires a solid understanding of how AI systems operate. One such job asked for someone “to design and optimize chatbot prompts that simulate realistic sales conversations aligned with client-specific sales processes and objectives.” Another job ad specified that candidates should have a “PhD in Computer Science, Artificial Intelligence, Machine Learning, NLP, or a related field”. Often, such roles require Python proficiency. 

However, prompting AI, in principle, is not usually that hard if you follow certain rules. First, you need to know the objective – for Info-pros, it’s likely to be:

• To research a topic – increasingly easy with AI deep research options; 
• To summarize material, sense-check text, or analyse data 
• To check material for spelling or grammar
• To draw images or possibly even write material. (If the aim is to publish, be careful – it’s NOT your own work, and so using AI for such purposes is problematic. However, I’ve used AI to write a legal letter, saving both time and money. I checked with a lawyer friend, and she said that it was pretty much what she’d have written!)

Next – be careful about how much information you give to the AI. Is bias OK? Do you want to lead the AI to give you the response you actually want, or that the AI thinks you want, from the way you asked? Although it’s a good idea to set the scene by saying who you are, what you want and in what format and why, you may also prompt the AI to give a response that is biased accordingly. 

As a trivial example, I asked ChatGPT about philosopher and author Ayn Rand: 

I’m a voting US Democrat; my leanings are towards AOC. Would books by Ayn Rand be worth reading?

ChatGPT said (in summary): “Given your political leanings, Ayn Rand’s books likely won’t align with your values … .” ; “But they may be worth reading to understand ideological opponents  .” “Rand’s works are often polemical and not nuanced.”

I got a completely different answer when I asked:  I’m a voting US Republican and I hate AOC. Would books by Ayn Rand be worth reading?

ChatGPT responded: “Yes, Ayn Rand’s books might be worth reading for you – especially given your political leanings …” In this case, there was no follow-on critique of Rand’s books. 

The neutral question I’m a US voter. Would books by Ayn Rand be worth reading? gave a much more nuanced answer, including reasons to read (if interested in limited government, free-market economics, or political foundations of modern conservatism/libertarianism) and downsides. “Her writing can be long-winded and didactic” was not brought up in the response to the Republican question, and another critique not mentioned in any response has noted that Rand was “an atheist and strongly anti-religion, which may clash with more traditional conservative values.”

To see the full answers: 

• Democrat: https://chatgpt.com/share/6852b754-9274-8001-9e63-fa4fb8e39e6e (plus an additional question on reading Karl Marx). 
• Republican: https://chatgpt.com/share/6852b729-20e8-8001-bfbc-4c05cd2180f0
• Neutral: https://chatgpt.com/share/6852b5be-c634-8001-9a7c-85acee7a1207

In my next blog in this series, I’ll go into more detail on what you should consider when prompting AI – so that you get what YOU need and not what the AI thinks you need.  

Arthur Weiss has been an infopreneur for almost 30 years. He founded AWARE in 1995 after a career at the business information company Dun & Bradstreet. He specializes in competitive and marketing intelligence using open sources (OSINT). Recently, he has pivoted to new areas, including exploring how AI tools can support infopreneurs. His latest insights can be read in International Marketing & Competitive Intelligence and Computers in Libraries magazines. He may be contacted at a.weiss@aware.co.uk.

Arthur will be hosting his next “AI Exchange” virtual event for AIIP on Wed, Sep 10, 2025 11:00 AM CDT. Register here!