ChatGPT can be very useful, but its default approach to questions often feels like a very boring textbook. This cautious approach to basic prompts makes sense on the one hand, as it tries to appeal to the widest possible audience. On the other hand, it can mean that you\u2019ll be disappointed by its answers if you want something beyond generic sentences.<\/p>\n\n\n\n
So let's take a look at some creative and simple phrases that you can insert into almost any prompt and consequently get much better use out of ChatGPT. Here are some of the best prompt formats for getting more engaging ChatGPT responses.<\/p>\n\n\n\n
ChatGPT's empty and somewhat encyclopedic tone is fine, but if you want something more, if you want it to be more actively helpful, you can assign it a role during the conversation. Suddenly, you have practically endless options for what roles you want ChatGPT to take. <\/p>\n\n\n\n
For example, you can ask ChatGPT to write you a list of negotiation tricks. The list of strategies and tactics it offers you can be helpful, but they almost certainly won't make you feel like you're ready for a real conversation.<\/p>\n\n\n\n
Instead, you can use the following prompt: \u201cLet's role-play a realistic situation where I'm negotiating for a higher salary and you're my boss, who has some concerns about the idea, which means we need to react appropriately.<\/em>“.<\/p>\n\n\n\n ChatGPT can start with: \u201cAs your manager, I appreciate your hard work, but budgets are tight this quarter. Why do you think now is the right time for a raise?<\/em>"Now you can actually practice what you're going to say and ChatGPT will continue the discussion with you from your boss's perspective. It may not anticipate everything in the conversation, but you'll likely be much better at answering those specific questions than you would otherwise be."<\/p>\n\n\n\n ChatGPT is a machine learning model, not a mind reader. If you talk to it regularly, it will remember the context of your conversations. It has a memory that will automatically add to it during your chats, and you can customize it as you wish. Why not take advantage of that?<\/p>\n\n\n\n A common question we ask ourselves is: "Which book should I read next?<\/em>\u201d and you\u2019ll get some good recommendations, like 1984 or To Kill a Mockingbird. The fact is, these are generic hits or classics that aren\u2019t tailored to you.<\/p>\n\n\n\n Check ChatGPT's memory or anything you may have said about yourself in the current conversation and use "Considering everything you know about me<\/em>\u201d and the answer will be very different. You can even include it in the prompt if you want.<\/p>\n\n\n\n For example: \u201cGiven everything you know about me, like my love of science fiction and my need for a book that will torment me emotionally, what should I read next?<\/em>“.<\/p>\n\n\n\n If you want ChatGPT to become an expert, you have to ask it. For example, if we ask it what it thinks about the future of artificial intelligence, at best it will offer us the pros and cons and maybe some pitfalls that it has identified. We can expect much better responses if we put ChatGPT in the role of an expert in a certain field. "Tell me what a data engineer would say about the future of artificial intelligence<\/em>", is a way to get a more in-depth response. To spice things up even more, we can confront opposing opinions: "Compare this with what<\/em> What would a Silicon Valley financier say about artificial intelligence?<\/em>". The conflict of opinions is much more engaging, although ChatGPT usually uses a cautious tone in its responses. <\/p>\n\n\n\n ChatGPT sometimes assumes you know more than you actually do and bombards you with jargon without context. Slowing it down a bit can make a big difference. You can ask it: \u201cHow does machine learning work?<\/em>"You'll probably end up with a very cluttered paragraph full of technical terms. Of course, if you don't understand them, you can quickly miss the point."<\/p>\n\n\n\n For the uninitiated, a better way is to ask ChatGPT to "explains how machine learning works step by step, as if you were teaching someone who has never coded before.<\/em>"No more jargon in this answer:"Imagine you want to teach a dog tricks. You show it how to sit, and every time it does, you give it a treat. Over time, the dog learns that \u2018sit\u2019 means good things are happening. Machine learning works in a similar way, except instead of dogs, you have data, and instead of treats, you have accuracy.<\/em>"Machines like dogs is the kind of analogy that many more people have understood."<\/p>\n\n\n\n Easier to understand doesn't necessarily mean simpler. You might just need the right person to explain it to you. Combine a simplified prompt with the idea of role-playing and you can get a much more entertaining response. Try asking ChatGPT to "Explain machine learning as if you were a stand-up comedian performing for an audience with no scientific background.<\/em>"Or"Explain machine learning like you're a pirate telling your crew about a mysterious new kind of treasure.\u201d<\/em><\/p>\n\n\n\n Ask a normal question and you'll get a normal answer. Ask something unexpected and things will get interesting. You can ask ChatGPT: "Tell me about the impact of social media on society.<\/em>". You'll get a useful, if predictable, breakdown of positive connections, technological advances, and entertainment opportunities, along with the rapid spread of misinformation, negative impacts on mental health, and authoritarian control of information. It sounds good, but it's nothing new.<\/p>\n\n\n\n Instead, ask ChatGPT: \u201cGive me an unconventional perspective on the impact of social media,<\/em>\u201d and you will find out that \u201cAlthough social media is often criticized for shortening attention spans, it may actually be enhancing microlearning. Users absorb complex ideas in small formats, making knowledge more accessible than ever. Unfortunately, this also includes malicious misinformation and corporate propaganda.<\/em>“<\/p>","protected":false},"excerpt":{"rendered":" ChatGPT can be very useful, but its default approach to questions often feels like a very boring textbook. This cautious approach to basic prompts makes sense on the one hand, as it tries to appeal to the widest possible audience. On the other hand, it can mean that you\u2019ll be disappointed by its answers if you want something beyond the generic [\u2026]<\/p>","protected":false},"author":2,"featured_media":0,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[12],"tags":[192],"class_list":["post-6086","post","type-post","status-publish","format-standard","hentry","category-triki-nasveti","tag-umetna-inteligenca"],"acf":{"subtitle":"Pri\u0161li smo do to\u010dke, ko je pri uporabi generativnih modelov umetne inteligence precej te\u017eko izstopati. Morda pa bodo ravno spodnji pozivi tisti, ki vam bodo pri tem pomagali.","heading":"","summary":"Pri\u0161li smo do to\u010dke, ko je pri uporabi generativnih modelov umetne inteligence precej te\u017eko izstopati. Morda pa bodo ravno spodnji prompti tisti, ki vam bodo pri tem pomagali.","thumbnail_small":"https:\/\/racunalniske-novice.com\/wp-content\/uploads\/2025\/03\/solen-feyissa-Aj7cDaR6QXs-unsplash-560x315.jpg","thumbnail_large":"https:\/\/racunalniske-novice.com\/wp-content\/uploads\/2025\/03\/solen-feyissa-Aj7cDaR6QXs-unsplash-1024x683.jpg","thumbnail_caption":"","gallery":"","video_gallery":null,"author":"","links":null,"sources":[{"title":"Tech Radar","url":"https:\/\/www.techradar.com\/computing\/artificial-intelligence\/5-better-prompts-to-use-with-chatgpt"},{"title":"Unsplash","url":"https:\/\/unsplash.com\/photos\/a-person-holding-a-cell-phone-in-their-hand-Aj7cDaR6QXs"}],"skip_language":[]},"yoast_head":"\n2. \u201cConsidering everything you know about me\u2026\u201d<\/h2>\n\n\n\n
3. \u201cWhat would an expert say about this?\u201d<\/h2>\n\n\n\n
4. \u201cBreak it down into steps for a beginner\u2026\u201d<\/h2>\n\n\n\n
5. \u201cGive me an unconventional take on\u2026\u201d<\/h2>\n\n\n\n