Photo by www.edweek.org

Everyone is discussing where AI should be used; in my opinion, we should also consider where it shouldn’t be.

Although AI is powerful, that doesn’t mean it is always an advantage to use it in every case. In certain situations, handing the task over to AI can eliminate the human judgment, responsibility, empathy, or experience that makes the task valuable in the first place.

Here are some places where I would be very careful, or avoid using AI altogether:

1. Helping you to make major life decisions. While AI can help you consider your options, it shouldn’t decide whether you should marry someone, end a relationship, move to another country, quit your job, or make any other deeply personal decision.
2. AI can help you organize your thoughts when you need to apologize, give feedback, settle a conflict, or let someone know how you feel; however, you still have to carry out the conversation.
3. It is impossible to judge a person’s character based on limited information. If you give an AI several messages and then ask it, “Is this person toxic?” or “Is my partner a narcissist?”, you will end up with a false sense of certainty about complicated human behavior.
4. It is not advisable to make medical decisions without the involvement of a qualified professional. While AI can help explain medical terms or prepare questions for your doctor, diagnosis, treatment changes, medication decisions, and emergency care should not be left to a chatbot.
5. Making legal decisions without proper scrutiny is one thing; using AI to understand a contract is something else, but relying on it alone for a lawsuit, an immigration matter, a tax problem, or a legally binding agreement is quite different.
6. People should not be assessed in high-stakes situations without human supervision. Matters such as hiring, firing, promotion, lending, insurance, education, and other decisions that have a major impact on someone’s life must not be left to AI’s judgment as a black-box system.
7. Every personal message is written. If AI produces all the birthday messages, condolence notes, thank-you notes, messages of love, and apologies, communication might become efficient but emotionally empty.
8. Replacing learning with answer generation; asking the AI for the final answer each time may make you faster now but will make you weaker later on. Sometimes, real learning comes from struggling with the problem.
9. Use it only when confidentiality matters; sensitive company information, private conversations, customer data, passwords, medical records, legal documents, or personal information must not be pasted into an AI system without first understanding how the data will be handled.
10. Avoiding responsibility through AI. The fact that ‘the AI told me to do it’ does not constitute accountability. When a decision affects other people, the person who makes it is still responsible for the outcome.
11. About showing empathy. In cases involving grief, serious illness, trauma, or a personal crisis, AI can help you choose the right words. Still, people generally need a real human presence rather than a perfectly worded reply.
12. Replacing your own sense of taste and judgment altogether. If the AI decides for you what you write, what you watch, what you buy, what you believe, what you create, and how you communicate, you will eventually become very skilled at producing things even though you don’t know what you actually think.

I am highly optimistic about AI, but I don’t believe the aim should be to apply it in every aspect of life.

Perhaps we should ask: “Does AI enhance this task, or does it take away something important from it?”

The wisest application of AI is knowing when it’s appropriate not to use it.