Top 5 misconceptions about AI

The use of artificial intelligence continues to grow. ChatGPT alone uses around 122 million users. Yet we still encounter inconsistencies and assumptions that surround the whole world of AI.

In this article we will try to get to the bottom of it. What do we think are the biggest misconceptions people associate with AI?

1. Artificial Intelligence = ChatGPT and other LLMs

Although most users most often encounter AI through ChatGPT and other language models (LLMs), this is just one of many types of AI.

LLMs are just a small slice of the vast field of AI. Talking about them as AI is like saying the internet = email.

In reality, AI encompasses a much broader range of technologies - and you've probably already encountered many of them. Here's a selection of them:

Machine Learning

It allows computers to learn from data - without having to be programmed step by step. Models learn to recognize patterns, relationships and rules that are not visible in the data at first glance, so they can make predictions, classify or make decisions even over new information.

Examples: disease prediction, risk assessment in banking, customer data analysis, weather forecasting

Computer Vision

Recognises and analyses images and videos.

Examples: face detection, license plate recognition, X-ray analysis, OCR (text recognition from image)

Reinforcement Learning

Reinforcement Learning (RL) is machine learning that learns optimal behavior through trial and error and the acquisition of rewards or punishments in a dynamic environment.

Examples: AlphaGo, robot control, AI in games, traffic control

Generative Models

They create new content.

Examples: DALL-E (pictures), Sora (video), Suno (music), ElevenLabs (voice), Veo3 (music)

Recommender Systems

They recommend content based on user behaviour.

Examples: Netflix, Spotify, YouTube, e-shops and even social networks such as Facebook or Instagram

2. AI = search engine

Many people use AI in a similar way to Google - they ask a question, expect an answer. But this leads to the mistaken impression that AI is just a "smarter search engine". In reality, they are completely different technologies.

A search engine (e.g. Google) searches the internet in almost real time. In contrast, most AI models (e.g. ChatGPT) generate answers from pre-learned data - and do not "look up" the actual information in the baseline. Some versions of AI (e.g., GPT-4o with browsing functionality) do have online access, but this is not automatic or universal.

Another difference is the energy intensity. Depending on the model, AI response generation is orders of magnitude more computationally intensive and energy consumption than a normal search.

Today, the distinction between AI and search is starting to blur - for example, Google is integrating Gemini directly into search results. Still, it's a combination of two different technologies.

3. Automation = AI

Recently, a lot of classical automation has been confused with the use of AI. But automation and AI are not the same - although they do occasionally interconnect. For example, you don't need AI to simplify a lot of routine tasks, but you can get by with automation. Tools like Make or Zapier can help you with those.

Automation means that the system performs predefined actions without human intervention - according to fixed rules. It doesn't think, it doesn't learn, it just repeats what it is told. An example is the order confirmation email when you buy something. When A happens, do B.

Artificial Intelligence on the other hand, learns from data, adapts to new situations and can make decisions that were not precisely programmed in advance. An example might be an email confirmation, but the content is tailored to your behaviour - for example, the AI will select tailored product recommendations.

4. Machine learning = AI = deep learning

These terms are often confused:

• Artificial Intelligence (AI) is the broadest term - it includes systems designed to solve problems and automate tasks that typically require human intelligence.
• Machine Learning is a subset of AI - these are techniques where models are learned from knowledge (data) instead of being manually programmed.
• Deep Learning is a subset of machine learning - it uses deep neural networks to automatically recognize complex patterns in data.

5. AI has its own opinion

When the AI responds fluently and confidently, it can appear to be expressing an opinion. But in reality no opinion - has no consciousness, no values, no convictions.

AI responses are the result of a calculation that predicts what words are statistically the best fit - does not express an intention, attitude or personal view.

The fact that something sounds human doesn't mean it is.

Conclusion

Artificial intelligence is still riddled with many inconsistencies - and there are many more than we have listed here. It is an extremely capable technology that we are still learning to use. Much like the internet used to be.

It has great potential, but it also has its limitations - and it is good to know them. That's why it's important to understand what AI really is, what it can do, what it can't do... and what we might think about it.

---

Related articles:

• TL;DR of the most commonly used AI terms - Getting lost in the terminology of the AI world? Then this article is for you.
• AI: Assistant or threat to juniors? - AI in development through the eyes of a junior: a valuable helper or an invisible crutch that hinders growth?
• How to know when the time is right to implement AI? - A practical guide for companies that are considering using AI but are still hesitant.
• The ecological impact of AI: What's going on behind the scenes? - Using AI is not without impact - every prompt triggers processes that consume both electricity and water.