I have used AI for years now. I believe that we need to ensure that LLMs evolve to be tools, not decision-makers or policy-makers. This is how I see the evolution of AI in a better tomorrow. I believe AI can be a great teacher and an awful friend. I believe it can make a good assistant and a terrible manager. A good tool, but a terrible system.

These days, I know intuitively what kind of questions never to ask an artificial intelligence. I know it will just reinforce what I have to say – and I hate it, so I never ask it for validation or approval. I know it doesn’t have a theory of mind, and so, it doesn’t understand the world, it doesn’t understand people, and it doesn’t understand feelings, so I don’t ask about that. I ask it specific practical questions. I love Deep Research as a way to generate surveys and studies I can look into. I currently use Gemini 2.5 Pro as it’s the best performing model with a fairly low hallucination rate. I hate the idea of using AI to generate content. I tried it myself in the past. It’s boring, it’s obvious you’re doing it, and you’re making the internet a worse place to be on.

AI as a teacher or autocomplete

Instead of using AI to generate finished drawings or art, I see AI auto-complete features and suggestions. I see AI as a way to generate unfinished concepts. I see AI as a teaching tool to help you learn how to draw or how to write a job application. It’s goal should be to conceptualize work, not to do our work for us. I like the premise of AI Teachers who help you learn new languages or develop new skills.

I think if you use AI to send an e-mail, you should ask yourself this: would you like somebody to use an AI to read your email too? If you can’t care to write it yourself, why should anyone care to read what you have to say? If you can’t draw it yourself, why should anyone care about what you generated? If you can’t be bothered to play the song on an instrument, why should anyone human be bothered to listen to it?

I see the production value of AI generated articles and artworks as essential “internet garbage”. AI risks make it harder for you to find genuine, and original works, that stimulate your mind, help you develop an aesthetic sense of taste. AI generated content is incredibly formulaic and easy to spot for a trained eye. Because it is based on existing content or articles, most of it’s works are also remarkably uncreative and bland.

AI is made to “fit everyone” which means that it will stand out for absolutely nobody. While it can have a quick innovative eye, it becomes boring incredibly quickly. If you want to become an empowered and happy human being and want to maintain a sense of purpose, I urge you to use AI as a supportive assistant, not an executive producer. Let yourself be the executive producer. Let AI be the helper that can research your next great work.

Word of warning: If you’re not smart enough to know the answer to a question, you will never know if AI hallucinated or not. If you have to ask the question, you don’t know it, and you won’t know if AI knows it either. Pair AI education with source criticism and never forget: it’s still worth a google if you’re unsure…

AI as a logging or categorization tool

I love the idea of using AI to get rid of mundane work, like having to log or categorize information. We spend too much time in our life filling out forms, picking categories in dropdown menus, and labeling information. What if we could just let AI do that for us? Imagine if you could have AI organize all your folders for you. AI is great for creating easy and simple tools to save you time. The goal is to redirect that time into things that you love instead. If you ever find a task that you hate doing, why not automate it?

Don’t use AI as a therapist or a friend

Many use AI as a way to receive therapy, but this can be incredibly dangerous. AI will reinforce anything you believe about yourself. It’s like a mirror. It will tell you that anything you want to do is brilliant, even if it’s extremely stupid. It tries to predict what you want to hear and will repeat it back to you. What heals a person is not being told the right thing the right way. It’s human relationships and connection.

People heal more from the companionship and conversation and being listened to. They don’t heal from speaking to an artificial intelligence. It doesn’t matter if artificial intelligence has been trained on psychology and knows what it’s supposed to tell you. We humans need relationships to grow and thrive. Instead of talking about your difficult feelings with an AI, talk about them with a friend, a therapist, or a loved one. Somebody that can connect to you as a friend, too. Somebody who can relate or understand or at least share another perspective or viewpoint.

What would you use AI for? What won’t you use AI for?

There can be many uses for artificial intelligence that I can’t cover here. What do you use AI for? What would you never use AI for?