How to properly use AI?

Just as the internet did decades ago, AI has quickly made its way into both markets and everyday life. On the business side, companies in the S&P 500, the Dow Jones index, and especially those in the Nasdaq, are investing billions of dollars in the technology. However, the return on investment has yet to materialize fully.

In daily life, the primary concern is that AI will replace human workers. So far, however, this is not happening. Many tech companies, including Amazon, have recently accelerated layoffs, but this may be due to broader concerns about a potential slowdown in the U.S. economy rather than AI itself.

So why do we need AI if it's not to replace people?

It is possible that artificial intelligence will surpass humans in certain jobs. However, for now, even Google CEO Sundar Pichai has recently advised users to maintain a critical attitude toward AI-generated responses, as large models are prone to errors, also known as AI “hallucinations.”

A recent research paper from OpenAI explains that these errors occur because models are designed to respond, even if it is not entirely correct. Researchers are teaching models to say “I don't know” when appropriate, but this still leaves open the question of the reliability of the underlying data.

For now, AI seems less like a replacement for humans and more like a powerful assistant, similar to how Excel once transformed the work of analysts and accountants. But like any tool, AI only works well if you know how to use it. Blindly asking it to “do the work for you” is rarely effective.

AI performs best when provided with clear context, a specific function, and a defined format for its output. It can make mistakes or invent details, so it's best to provide the exact text or source you want it to use, such as a law or official document, and always manually revise the response.

What would happen if the “AI bubble” everyone is talking about suddenly burst and companies cut back on AI spending? Like the internet, AI is not going away. It will remain with us and continue to evolve. That's why now is a good time to learn how to use it to your advantage.

Finally, be careful with sensitive data when using AI tools. The lawsuit in the United States accusing Google of tracking users through its Gemini model suggests that other companies may also be handling private information in ways we don't fully understand.