{"id":67514,"date":"2025-06-19T08:11:29","date_gmt":"2025-06-19T08:11:29","guid":{"rendered":"https:\/\/morningtrain.dk\/what-is-an-ai-model\/"},"modified":"2026-04-17T12:05:52","modified_gmt":"2026-04-17T12:05:52","slug":"what-is-an-ai-model","status":"publish","type":"post","link":"https:\/\/morningtrain.dk\/en\/what-is-an-ai-model\/","title":{"rendered":"What is an AI model?"},"content":{"rendered":"\n

AI<\/p>

19. juni 2025<\/p>

What is an AI model?<\/h1><\/div>

In its simplest form, an AI model is a set of rules and patterns that a computer has learned, typically using large amounts of data. It is a bit like a recipe: it learns to recognise what typically follows something else, and that is how you can \u201ctrain\u201d the AI. <\/p><\/div>\n\n

\"\"<\/figure><\/div><\/section>\n\n
\n
\n

\n Indhold\n <\/p>\n \n Introduktion\n <\/a>\n \n Eksempler p\u00e5 modeller\n <\/a>\n \n Hvad kan de bruges til\n <\/a>\n \n Fordele og ulemper\n <\/a>\n \n S\u00e5dan kommer du i gang\n <\/a>\n \n Beskrivelse af modeller\n <\/a>\n \n FAQ\n <\/a>\n <\/div>\n

\n \n\n

An AI model is a recipe that learns from data: it helps with text, images, analysis, and much more. You get the most value by starting simple\u2014use ready-made tools, ensure human oversight, and work iteratively with your colleagues. <\/p>\n\n\n\n

When you train an AI (most often using large amounts of data), it learns to recognise what typically follows something else.<\/p>\n\n\n\n

For example, a model can learn that \u201chorse\u201d is often followed by \u201cgallop\u201d in a text, or how an image of a cat differs from a dog. Once the model has practised enough, it can predict, provide answers, or create something creative based on what it has learned. <\/p>\n\n\n\n

Examples of AI models<\/h2>\n\n\n\n

Here are some well-known models:<\/p>\n\n\n\n

Model\/tool<\/th>Developer<\/th>What it is good for<\/th><\/tr><\/thead>
ChatGPT<\/a><\/strong><\/td>OpenAI<\/td>Conversations, text, ideation, brainstorming, explanations<\/td><\/tr>
Claude<\/a><\/strong><\/td>Anthropic<\/td>Long texts, empathetic language, collaboration, safety and ethics<\/td><\/tr>
Gemini<\/a><\/strong><\/td>Google DeepMind<\/td>Multimodal AI (text, image, audio), search, coding, analysis<\/td><\/tr>
DALL\u00b7E<\/a><\/strong><\/td>OpenAI<\/td>Creating images from text\u2014creative and illustrative<\/td><\/tr>
Midjourney<\/a><\/strong><\/td>Independent team<\/td>Aesthetic, atmospheric images, art and branding<\/td><\/tr>
Stable Diffusion<\/a><\/strong><\/td>Stability AI<\/td>Open-source image model, extensive visual control, local deployment<\/td><\/tr>
TensorFlow<\/a><\/strong><\/td>Google<\/td>Train and run AI models, image recognition, production and scalability<\/td><\/tr>
PyTorch<\/a><\/strong><\/td>Meta<\/td>Flexible AI development and research, text and image models<\/td><\/tr>
GitHub Copilot<\/a><\/strong><\/td>Github + OpenAI<\/td>Helps with coding directly in the editor, auto-suggestions and code understanding<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Read more about each model further down, or by clicking the model name.<\/p>\n\n\n\n

You can read more about our view on the use of AI, which tools we use and what for, in this post<\/span><\/a>.<\/p>\n\n\n\n

What can you use an AI model for?<\/h2>\n\n\n\n

Chat and customer service<\/strong>An AI model can live in a chat on your website and help with customer questions\u201424\/7, without breaks.<\/p>\n\n\n\n

Text generation and writing support<\/strong>Do you ever think: \u201cHow do I start this email?\u201d\u2014an AI model can suggest an opening, a headline, or summarise texts for you.<\/p>\n\n\n\n

Image and pattern recognition<\/strong>For businesses with images: models can sort images, find defects, or recognise objects.<\/p>\n\n\n\n

Data analysis and predictions<\/strong>Do you need to forecast sales, inventory needs, or risk? AI models can help you identify patterns in the numbers. <\/p>\n\n\n\n

Automation<\/strong>Repetitive tasks\u2014e.g., reading forms, retrieving information, filling in fields\u2014can be delegated to AI models.<\/p>\n\n\n\n

Creative use<\/strong>Music, illustrations, scriptwriting\u2014models can help with ideas or even carry out part of the creative work.<\/p>\n\n\n\n

Pros and cons of using AI models<\/h2>\n\n\n\n
Advantages<\/th>Cons<\/th><\/tr><\/thead>
\u2705 Save time on routine tasks<\/td>\u274c Can fail or misunderstand\u2014requires oversight<\/td><\/tr>
\u2705 Scales\u2014responds 24\/7<\/td>\u274c Can be expensive to train or use at scale<\/td><\/tr>
\u2705 Supports creativity\u2014ideas and inspiration<\/td>\u274c Risk of bias\u2014the model learns from our data and can repeat mistakes<\/td><\/tr>
\u2705 Automates error-prone tasks<\/td>\u274c Can feel like a \u201cblack box\u201d\u2014you do not always know how it arrives at its answers<\/td><\/tr>
\u2705 Increases your capacity and therefore your competitiveness<\/td>\u274c AI requires significant server power and therefore has a large CO2 footprint<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Best practices: how to get started smartly<\/h2>\n\n\n\n
    \n
  1. Start simple<\/strong>Use ready-made tools (ChatGPT, Copilot, Zapier) rather than training your own. You will get value quickly\u2014without the hassle. <\/li>\n\n\n\n
  2. Data and responsibility<\/strong>Consider what data you use and who has access. Remember GDPR and data protection. <\/li>\n\n\n\n
  3. Critical and human oversight<\/strong>Never let the AI model stand alone\u2014an employee must check the output and quality so mistakes do not slip through.<\/li>\n\n\n\n
  4. Iterate and adapt<\/strong>Start with small pilot projects and evaluate: What works? What creates value? Then you can scale up quickly. <\/li>\n\n\n\n
  5. Build user trust<\/strong>Tell the user when it is AI: \u201cThis answer was generated with support from an AI model.\u201d Transparency creates confidence\u2014both internally and externally. <\/li>\n\n\n\n
  6. Follow-up and learning<\/strong>Gather feedback: What works best? What are the challenges? Use the experience to improve the process. <\/li>\n<\/ol>\n\n\n\n

    Description of different AI models<\/h2>\n\n\n\n

    ChatGPT (from OpenAI)<\/strong><\/h3>\n\n\n\n

    ChatGPT is a language model that understands and writes text. You can ask questions, get explanations, brainstorm ideas, get help with emails, coding, and much more\u2014all in a human conversational style. <\/p>\n\n\n\n

    What is it good for?<\/strong><\/p>\n\n\n\n