> For the complete documentation index, see [llms.txt](https://skillful-ai.gitbook.io/skillful-ai-public-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://skillful-ai.gitbook.io/skillful-ai-public-docs/skillfulai_docs_v1.0/introduction/ai-models.md).

# AI Models

Today, there’s no news agency that hasn’t been coveting the AI hype. This hype gained momentum when ChatGPT was launched, powered by the GPT-3.5 model.

To understand the excitement, it’s important to first know what an AI model is.

**So, what’s an AI model?**

Well, an AI model is basically a computer program that can look at data, find patterns, and make decisions on its own. 

It uses rules (algorithms) to understand information and perform tasks like understanding speech or recognizing pictures, similar to how humans think.

But these model needs to be “trained” - teaching an AI model how to do specific jobs by showing it lots of examples.

**But how do you create an AI model?** 

<figure><img src="https://3620906879-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Ft1TLw2CAY3Yjr54moPYE%2Fuploads%2FbAWZsPcyOJsqJztYBCBy%2Fimage.png?alt=media&amp;token=d1b2e2bd-81d0-4633-ab6c-c0f7a81851e9" alt=""><figcaption><p>How to Create an AI Model</p></figcaption></figure>

Creating an AI model involves a few key steps.

* Define the specific problem to be addressed, such as predicting trends or classifying data.
* Collect and prepare relevant data, including cleaning to remove inaccuracies and inconsistencies.
* Select an appropriate model type based on the nature of the problem, such as supervised learning for labelled data or deep learning for handling complex patterns. &#x20;
* Train the model using a subset of the data, enabling it to identify patterns and make predictions.
* Evaluate the model’s performance on a separate dataset to measure its accuracy and effectiveness. Apply optimization and fine-tuning to enhance its performance further.
* Deploy the validated model in real-world settings, with continuous monitoring to ensure it remains effective and adapts to new data or evolving conditions as needed.