Why can’t you make an artificial manager with ChatGPT?

Posted by

In a previous blog post, I described how ChatGPT works and what the ability to analyze and then synthesize text gives us. I summed up the entry by saying that you can’t make a robot manager from ChatGPT, because this mechanism still doesn’t know what a human manager actually does, and actually has no way to find out based on encyclopedic knowledge. Today I will write more about it.

Despite the fact that the capabilities of the GPT-based chatbot amaze millions of people around the world (OpenAI’s application was the most popular app of all time in recent weeks), after a closer look at how this mechanism works, I don’t see how it can be used to turn a human manager into an artificial manager in team management. Why? There are two reasons.

First, the ontology of the world that ChatGPT uses is completely scattered and inconsistent. When we ask ChatGPT how to make scrambled eggs, we get a reasonably adequate description of how to do it, what we need and what comes out of it. But this is a problem well described in any literature, from cookbooks to cooking blogs. Its solution has some structure and chronology – you can’t crack an egg only after mixing it in a pan. To solve it, we also need familiar physical objects, such as eggs, a frying pan, fat, etc. Virtually always the same – the ontology of making scrambled eggs is fixed, known and repeatable. No new entities are created in the time that passes from one scrambled egg to another. There is always a need for eggs, a pan and so on….

In the case of management, it is quite different. Let’s ask ChatGPT how to set a goal. Here’s what we get in response.


To set a goal, go through several steps:

Think about what you want to achieve – think about what is important to you and what you want to achieve. You can start with questions like “What do I want to achieve?”, “What is my life goal?”, “What do I want to experience?”, etc.

Clarify your goal – once you know what you want, clarify your goals so that they are measurable and achievable. A good way to do this is to use the SMART method, which means formulating goals that are:

  • Specific (specified)
  • Measurable
  • Achievable
  • Relevant
  • Time-bound (time-specific)

Find your motivation – when setting goals, it’s important to find the motivation that helps you achieve them. You may want to think about what motivates you and what may discourage you from achieving your goals.

Create an action plan – in order to achieve your goal, you need to create an action plan to achieve your goal. In the plan, you can include the steps you need to take to achieve the goal and set deadlines for each step.

Monitor your progress – it is important to monitor your progress and regularly evaluate whether your actions are producing the desired results. This way you can adjust your actions if something is not working as planned.

Summarize and celebrate – when you reach your goal, it is important to take stock of your achievements and celebrate your success. This will help motivate you to achieve more goals.

We’ll get confusion with confusion – the SMART principle (and that’s the only sensible part about it), something about motivation, the need to create a plan, monitoring progress (sounds like 4 functions of management). Can any of you do this? Do any of you at least understand what you need to do to set a goal?

This is such an unspecific method of goal setting, described only in buzzwords, that you don’t even know how to take the first step to the goal of setting some kind of goal! The ontology of organizational reality is scattered and incoherent, which means that ChatGPT cannot create a coherent vocabulary and tell the story in a concrete, repeatable way.

Second, knowledge about management, such as a team or a project, is told in books and websites from the point of view of past situations or good advice. The problem is that the description of these situations usually (or even never!) does not include an inventory of the managerial actions that were performed to solve some organizational problem. Such a description would never sell and such a story would be as boring as a recipe for making a pencil. Who would buy such a management book or read such a blog (and after all, this is where ChatGPT gets its knowledge from)? As for the list of advice – such advice is usually quite general, follows from a rational view of how to do or achieve something and… has little to do with complex reality. To sum up – ChatGPT has nowhere to learn how to manage a team or project, i.e. how to be a manager.

In conclusion, because of the lack of a coherent ontology of organizational reality, spread in books and on the Internet, and the real knowledge of what a manager really does, gathered from these sources, it is impossible to make an artificial manager from ChatGPT. ChatGPT knows something, but inaccurately and poorly. A bit like a five-year-old child who is supposed to manage an expedition to Mars.

Finally, two interesting facts.

The first – I asked ChatGPT how to make an artificial manager. Here is his answer. Judge for yourself if he knows how to make an artificial manager….

Me: How to make a robot manager?


An artificial manager is an artificial intelligence-based tool that can help manage projects, business processes and human resources. Here are some steps you can take to create an artificial manager:

1. Define the goals and business objectives you want to achieve – the manager robot must be programmed with an understanding of the organization’s goals and objectives.

2. Define the tasks and processes to be managed – you need to specify the tasks and processes to be performed by the artificial manager, such as managing project schedules, coordinating team activities, monitoring work performance, etc.

3. Choose the right tools and technologies – the artificial manager can be built using a variety of tools and technologies, including artificial intelligence, machine learning, data analytics and robotics.

4. Configure artificial intelligence algorithms and models – the manager robot requires a range of artificial intelligence algorithms and models that will be responsible for analyzing data, making decisions and performing tasks.

5. test and improve your artificial manager – during the development of the robot manager, you should regularly test its performance and implement improvements to ensure that the artificial manager performs as expected and is effective.

It is important to remember that the artificial manager can be a useful support tool, but it will not replace a human manager, who will still be needed to make strategic decisions and manage the team.

The second – the story of a company that was founded by a human on March 25, 2023, but is led by ChatGPT. Whether this is a credible story, judge for yourself. Or is it the work of ChatGPT?