Become a TechRadar Insider
As the furore over losing ChatGPT-4o has shown, people don’t like it when AI presents robotic outputs, overly generic answers, or starts to sound stiff and defensive. I’ve found that the solution to avoiding this with later ChatGPT models is to use the principles from Dale Carnegie's How to Win Friends and Influence People because it has dramatically improved the quality of responses I get from ChatGPT.
Carnegie classic 1936 book reduces human interaction down into a set of practical social principles. A lot of modern management training, sales advice, networking culture traces back to them. If I were to boil the book down to a simple piece of advice, it would be that people respond better to encouragement and self-interest than blunt correction.
The strange thing is that Carnegie’s principles work perfectly on ChatGPT, because modern AI has been trained so heavily on human conversational expectations. That means that advice written for human relationships in 1936 now improves interactions with machines today.
Fundamental techniques in handling people, and AI
Carnegie suggests that we don't criticize, condemn, or complain, and give honest and sincere appreciation so we can arouse in the other person an eager desire to be helpful. Let’s look at five of Carnegie’s principles, one by one, and see how we can apply them to AI chatbots. I’ll give an example of a weak and better prompt for each, where I can.
1. “Begin in a friendly way”
In terms of AI, this means we need to give emotional context and intent before we demand it starts a task.
Weak prompt: “Rewrite this email”.
Better prompt: “I’m trying to sound warm and professional without sounding stiff. Can you help me rewrite this email?”
The result of doing it this way is often better tone matching, less robotic phrasing and fewer weird over-corrections. It turns out that ChatGPT performs better when it understands why you’re asking, not just what you want.
2. “Talk in terms of the other person’s interests”
This is probably the strongest principle for AI prompt engineering, and it means you need to explain the end goal before you ask for a response.
Weak prompt: “Summarize this document.”
Better prompt: ”Summarize this document so I can turn it into a short email for non-technical readers.”
Adding the end goal will result in more useful summaries, strong framing and less generic filler from ChatGPT. Think of it this way — the best prompts are more like briefings to a work colleague than commands to a robot.
3. “Be a good listener”
Most people are using ChatGPT backwards — they’re rushing straight to answers instead of turning it into more of a conversation. This one works brilliantly with follow-up prompting. Instead of dumping instructions, let the AI ask clarifying questions.
Prompt: “Before answering, ask me three questions that would help improve the result.”
Try this more often and you’ll notice that you get dramatically more personal answers, less hallucinated assumptions and more usable results.
4. “Give honest and sincere appreciation”
The idea of giving appreciation to ChatGPT sounds ridiculous at first, but positive reinforcement helps refine outputs.
For example, replying to something that wasn’t quite right with: “That structure was really close to what I wanted. Can you keep that same tone but make it shorter and punchier?” Gets much better results than simply “Try again, but shorter”.
By telling ChatGPT what it is about the last response you actually liked and what it did well, you get more consistency and less drift between versions. Even though the AI doesn’t have feelings, conversational reinforcement still shapes the interaction.
5. “Try honestly to see things from the other person’s point of view”
Converting this principle into something you can use with ChatGPT means assigning perspective and audience to your prompts.
Weak prompt: ”Explain quantum computing.”
Better prompt: “Explain quantum computing from the perspective of someone who finds physics intimidating and usually avoids technical subjects.”
I’ve also seen versions of this where you ask ChatGPT to explain it as if you’re reassuring a confused friend, not lecturing a classroom. It results in clearer explanations, less jargon and a more human-feeling response.
Try these tips the next time you interact with your AI chatbot of choice. You’ll find that Dale Carnegie’s principles work because modern AI has effectively been trained to simulate cooperative human conversation. It’s a strange paradox that the more conversational AI becomes, the more old-fashioned social skills start working on machines too.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
