Danial Hooshyar: the real power of AI lies in understanding data

His work focuses on AI in education, learning analytics, educational data mining, and adaptive educational systems.

You started your work with AI in education. What insights about how AI works have you found to be useful across different sectors?

Generally speaking, AI can work in two main ways. Sometimes it analyzes data independently and finds patterns. Other times, we guide it to look for specific patterns based on our expectations. 

One important insight is that AI is not just for making predictions. It can also be used to learn from data and understand what is happening inside it. Instead of only training a model and to predict outcomes, we can also examine what patterns the model has learned from the data.

This is useful across many sectors because it helps organizations make data-driven decisions. For example, businesses, healthcare, or education can all use AI to understand trends in their data better. 

It’s also important to note that large language models (LLMs) are not designed to analyze your own data deeply. For that, standard machine learning methods are often needed.

How do you explain the core logic of AI systems so that non-technical professionals can understand how they actually work?

At a basic level, AI enables machines to perform tasks that normally require human thinking. Some AI systems are rule-based, meaning they follow fixed instructions defined by experts. Others use machine learning, where systems learn patterns from data instead of predefined rules.

More advanced deep learning systems can learn from images, text, and other complex data, but their decision-making processes are often difficult to interpret. Generative AI is the latest type and can create text, images, and more, although it also raises challenges related to bias and transparency.

Who are your AI trainings most relevant for in the public and private sectors?

These trainings are relevant for anyone interested in understanding and analyzing data, regardless of their background in statistics or computer science. Participants learn to apply AI methods through visual programming, with no coding required.

The trainings are useful for professionals working with surveys, digital trace data, or other forms of data in areas such as business, education, or the public sector. AI methods can help organizations discover patterns and relationships that traditional statistical approaches may overlook.

In your experience, what do people most often misunderstand about how AI actually works?

Many people misunderstand what AI really is. Some associate it only with robots, while others think only of tools like GPT. 

In reality, AI has been used for decades and is already part of everyday systems, such as (Google) search engines or industrial processes. AI is not just about automation or personalization – it is mainly about finding patterns in data, optimizing decisions, and helping people gain deeper insights.

How does understanding how AI works change the way professionals use AI tools in their work?

When professionals understand how AI works, they begin to use it more thoughtfully and effectively. Instead of relying only on ready-made tools, they can use AI methods to analyze their own data, discover complex patterns, and even make predictions. For example, people can use data from devices like smartwatches to better understand their own behavior and make more informed decisions.

They also know better which AI tools are suitable for their tasks and which ones they should avoid.

What is one technology trend that you think is overestimated, and one that is underestimated?

Large language model (LLM) tools in education are often overestimated. Although they are widely used for tutoring and feedback, research shows they still struggle to reliably model student learning over time, even with fine-tuning.

Simply improving LLMs does not solve these deeper limitations, because they are not designed with learning science in mind.

At the same time, AI methods for data analytics and data mining are often underestimated. The real power of AI lies in building machine learning models on real data to discover patterns and support decision-making.

How do you see the role of a specialist in your field evolving over the next 5–10 years?

Over the next 5–10 years, the role of a specialist in AI and data-related fields will shift from simply applying tools to deeply understanding and guiding how they are used. Specialists will need to continuously learn and adapt as new methods and technologies emerge, while also critically evaluating their limitations. 

Routine tasks will increasingly be automated, so the real value will lie in critical thinking, creativity, ethical judgment, and the ability to guide AI tools effectively. Their role will also involve designing reliable, data-driven solutions and helping organizations make informed decisions based on real insights.

What kind of mindset best supports people in adapting to technological change?

People should be open to using AI, but also approach it critically and responsibly. AI is becoming a permanent part of education and everyday work, yet it is important to understand its limitations, especially in areas involving learning and decision-making.

Users should ask how AI systems are built, what data they rely on, and whether their outputs are trustworthy, fair, and transparent. Critical thinking and responsible use are essential for ensuring that AI supports people effectively and safely.