The New Skill: “Let Me Check That.”

Every important technology arrives with a promise. It will make us faster. It will make us smarter. It will take care of the tedious things so we can devote ourselves to more meaningful work. Sometimes that promise is fulfilled. Sometimes it is not. Usually, the truth is somewhere in between.

During my career at IBM, I saw wave after wave of technology change how people worked. The personal computer moved computing from behind glass walls and into offices and homes. The internet changed how organizations communicated with customers. Email sped up the flow of information but also created an endless stream of it. Each breakthrough made certain tasks easier while creating new responsibilities for the people using it. Artificial intelligence, possibly going to be called super intelligence, is the newest and perhaps most dramatic version of the information systems evolution.

Today’s AI systems can draft letters, summarize meetings, explain complex subjects, write computer code, translate languages, create images, organize travel plans, and offer suggestions on almost anything. The newest systems can even take multiple steps toward a goal: find information, compare options, prepare a document, and revise it after receiving feedback. It is impressive, but it can also be extremely useful. AI has given us a new skill to develop: knowing when to say, “Let me check that.” That sentence may become one of the most valuable habits of our time.

AI is often astonishingly fluent. It gives answers in complete sentences, with confidence, detail, and a polished tone. That is precisely what makes it helpful and what can make it dangerous. A response can sound intelligent while containing an incorrect date, a made-up quotation, an outdated recommendation, or a conclusion that does not fit the situation.

The problem is not that AI is trying to deceive anyone. It has no intention in the human sense. It is doing what it was designed to do: recognize patterns and produce a likely response. But “likely” is not the same as “true and accurate.”

We have all seen a version of this in ordinary life. A friend tells us the date of an event. A website lists a restaurant as open when it has closed. A GPS sends us down a road that is blocked by construction. The information may have been reasonable when it was created, but the world has changed.

AI can make this problem more subtle because its answers are so conversational. It does not simply provide a list of links and say, “You decide.” It often gives us the feeling that someone knowledgeable has already done the thinking. That is when we need to pause.

Imagine asking an AI system about a medication. It may offer a clear explanation of uses, side effects, and interactions. That can be helpful preparation for a conversation with a doctor or pharmacist. Most people don’t know how or take the time to give your AI your complete medical history, your lab results, or your allergies. The right response is not, “The computer said this is fine.” The right response is, “That is useful information—let me check that with my doctor.”

Or imagine asking AI to plan a trip. It may quickly suggest flights, hotels, restaurants, museums, and driving times. It may even assemble a beautiful itinerary. But before booking, someone still needs to confirm the dates, the cancellation rules, the accessibility of the hotel, the current hours of the museum, and whether the restaurant is actually open. A marvelous travel plan is not much use if it gets one important detail wrong.

The same is true in business. AI can help someone write a proposal, analyze a spreadsheet, summarize a contract, or compose a speech. It may save hours of work but it cannot take responsibility for the result. A person must still ask: Are the numbers correct? Is the conclusion fair? What assumptions are hidden in this answer? What important fact has been left out? The person who signs the letter, approves the plan, advises the client, or makes the decision still owns the consequences.

That is why I think of AI as a brilliant junior colleague. It is fast, energetic, well-read, and capable of producing a first draft in moments. It may notice possibilities we missed. It may help us get started when we are staring at a blank page.

But we should not hand AI the keys to the company or to our judgment. A good junior colleague needs direction. We tell them what we are trying to accomplish. We give them background. We ask questions. We correct mistakes. We do not assume that a confident presentation is the same as deep understanding. AI deserves the same kind of supervision.

This is not a negative view of the technology. Quite the opposite. I believe the best use of AI is to make people more capable, not less thoughtful. Used well, it can reduce drudgery and give us more time for the things machines do not do well: listening, building trust, understanding context, making ethical choices, and caring about the outcome. Those are not small things. They are the heart of good work and good leadership.

The National Institute of Standards and Technology, which has developed guidance for trustworthy AI, emphasizes qualities such as reliability, safety, transparency, privacy, fairness, accountability, and human oversight. In other words, the experts understand the goal is not simply more powerful technology. The goal is technology people can use responsibly. Human oversight is not an admission of failure. It is a recognition of reality.

We do not stop using calculators because they can produce an answer even when we typed the input incorrectly. We learn to check whether the answer makes sense. We do not abandon spell-check because it sometimes changes the wrong word. We read what we are about to send. We do not reject GPS because it occasionally directs us poorly. We look at the road.

In a world where machines can generate words, images, recommendations, and plans almost instantly, the missing resource may not be information. We already have more information than any of us can absorb. The scarce resource is judgment.

The future will belong not to the people who blindly accept every answer from a machine, nor to those who refuse to use new tools out of fear. It will belong to those who combine curiosity with caution: people willing to explore, learn, test, and verify.

Before we act on something important, let us keep one phrase close at hand: “Let me check that.” It is not a rejection of intelligence, artificial or otherwise. It is the beginning of wisdom.