{"id":6664,"date":"2026-02-09T12:42:51","date_gmt":"2026-02-09T12:42:51","guid":{"rendered":"https:\/\/maxymia.com\/?p=6664"},"modified":"2026-02-19T12:46:31","modified_gmt":"2026-02-19T12:46:31","slug":"errores-al-escribir-prompts-y-como-evitarlos","status":"publish","type":"post","link":"https:\/\/maxymia.com\/en\/errores-al-escribir-prompts-y-como-evitarlos\/","title":{"rendered":"Errors when writing prompts in AI"},"content":{"rendered":"

Let's be honest: We've all written awful prompts at some point.<\/strong>. You ask ChatGPT, Gemini, or another AI a question, and the answer is nothing like what you expected. And you think: \u201cAI is useless!\u201d<\/em>.<\/p>\n\n\n\n

Spoiler alert: it's almost never the AI's fault. Usually, the problem lies with how are we asking<\/strong>.<\/p>\n\n\n\n

Writing prompts is like giving instructions in a lab: if they aren't clear, the result will be\u2026 unpredictable. Let's look at the most common mistakes and, most importantly, how to correct them<\/strong>.<\/p>\n\n\n\n

Errors when formulating a prompt in ChatGPT, Gemini, or other AI<\/h2>\n\n\n\n

1. Ask in a vague or ambiguous way<\/h3>\n\n\n\n

This is the classic one. The most common. The most human.<\/p>\n\n\n\n

Example of a problematic prompt: \u201cExplain this topic to me.\u201d What topic? At what level? With what objective?<\/p>\n\n\n\n

When the prompt is vague, the AI fill in the gaps as best you can<\/strong>. And sometimes he does it well\u2026 and other times not so much. How can it be improved? <\/strong>Be a little more specific, don't be afraid.<\/p>\n\n\n\n

\u201c"Explain this concept to me clearly, with simple examples and a scientific approach."\u201d<\/p>\n\n\n\n

You don't need to write a thesis, just give clear clues<\/strong>.<\/p>\n\n\n\n

2. Not defining context or objective<\/h3>\n\n\n\n

The AI doesn't know if you're writing an article, preparing a class, or just trying to understand something out of curiosity. If you don't tell it Why do you want the information?<\/strong>, It will give you a generic answer.<\/p>\n\n\n\n

Typical example:<\/p>\n\n\n\n

\u201c"Summarize this article." For whom? For what purpose? Very short or detailed?<\/p>\n\n\n\n

How can it be improved? <\/strong>Add context and purpose.<\/p>\n\n\n\n

\u201cSummarize this article highlighting the hypothesis, methodology, and main results, intended for a literature review.\u201d<\/p>\n\n\n\n

You'll see how the answer changes completely.<\/p>\n\n\n\n

3. Expecting AI to \u201cguess\u201d what you need<\/h3>\n\n\n\n

This mistake is very common, especially when you've been working with AI for a while and you think: \u201cHe should understand me by now\u201d<\/em>. But no. AI It has neither infinite contextual memory nor human intuition<\/strong>.<\/p>\n\n\n\n

Example:<\/p>\n\n\n\n

\u201cDo better.\u201d<\/p>\n\n\n\n

Better how? Shorter? More technical? Clearer?<\/p>\n\n\n\n

How can it be improved? <\/strong>Tell him exactly what you want to improve.<\/p>\n\n\n\n

\u201cMake the text clearer, eliminate redundancies, and maintain an academic tone.\u201d<\/p>\n\n\n\n

AI works best with specific instructions<\/strong>, not with assumptions.<\/p>\n\n\n\n

4. Asking for too much in a single prompt<\/h3>\n\n\n\n

Here we tend to err on the side of enthusiasm.<\/p>\n\n\n\n

Chaotic example:<\/p>\n\n\n\n

\u201cSummarize the article, translate it, draw conclusions, identify methodological errors, and propose future lines of research.\u201d<\/p>\n\n\n\n

Can he do it? Sometimes, yes. Will he do it well? Not always.<\/p>\n\n\n\n

When you ask for too many things at once, the AI tends to respond superficially<\/strong> or to leave parts lame.<\/p>\n\n\n\n

How can it be improved? <\/strong>Divide and conquer. First summarize. Then analyze. Then propose improvements.<\/p>\n\n\n\n

Use several chained prompts, just like you would do yourself.<\/p>\n\n\n\n

How to fix a poorly designed prompt<\/h2>\n\n\n\n

Do you have a prompt that isn't working? Don't throw it away. Refine it<\/strong>.<\/p>\n\n\n\n

Here's a short, practical method:<\/p>\n\n\n\n

    \n
  1. Detect what's wrong<\/strong>
    Is it too generic? Is there a lack of context? Is the answer unhelpful?<\/li>\n\n\n\n
  2. Add intention<\/strong>
    What exactly do you want to achieve with the answer?<\/li>\n\n\n\n
  3. Define format and tone<\/strong>
    List, summary, long text, technical level, target audience.<\/li>\n\n\n\n
  4. Reformulate without starting from scratch<\/strong>
    Sometimes all it takes is one extra line.<\/li>\n<\/ol>\n\n\n\n

    Example of evolution:<\/p>\n\n\n\n

    Initial prompt:<\/p>\n\n\n\n

    \u201cExplain this to me.\u201d<\/p>\n\n\n\n

    Prompt corrected:<\/p>\n\n\n\n

    \u201c"Explain this study to me clearly, highlighting the objective, methodology, and main results, with a focus on researchers."\u201d<\/p>\n\n\n\n

    The difference is enormous.<\/p>\n\n\n\n

    With a few tweaks, you'll go from mediocre answers to truly useful results. And remember, if AI is a key resource for you, you might be interested in this course on prompt engineering<\/a>.<\/p>","protected":false},"excerpt":{"rendered":"

    Let's be honest: we've all written terrible prompts at some point. You ask ChatGPT, Gemini, or another AI a question, and the answer is nothing like what you expected. And you think, "AI is useless!" Spoiler alert: it's almost never the AI's fault. Usually, the problem lies in how we're asking the question. Writing [\u2026]<\/p>","protected":false},"author":109,"featured_media":6666,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_kadence_starter_templates_imported_post":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-6664","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ciencia-y-tecnologia"],"blocksy_meta":[],"acf":[],"jetpack_featured_media_url":"https:\/\/maxymia.com\/wp-content\/uploads\/2026\/02\/errores-al-escribir-un-prompt.jpg","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/posts\/6664","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/users\/109"}],"replies":[{"embeddable":true,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/comments?post=6664"}],"version-history":[{"count":1,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/posts\/6664\/revisions"}],"predecessor-version":[{"id":6667,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/posts\/6664\/revisions\/6667"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/media\/6666"}],"wp:attachment":[{"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/media?parent=6664"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/categories?post=6664"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/maxymia.com\/en\/wp-json\/wp\/v2\/tags?post=6664"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}