Are We Becoming Too Dependent on Tech and AI?

Dependence on technology and AI becomes a problem when we can produce an answer but cannot explain it, check it, or handle the task when the tool disappears. Using a calculator, GPS, or an AI assistant does not automatically cross that line. Surrendering judgment deserves closer attention.

In a short clip from Nikhil Kamath’s podcast, he describes previously writing speeches himself and feeling less able to do so now. He also uses mental multiplication as an example of a skill he feels he has lost. His question: what happens to core human abilities when technology keeps removing the need to practise them?

The concern is reasonable. The conclusion that everyone is becoming less intelligent goes beyond the evidence. A useful answer starts with the particular skill, the way we use the tool, and what happens when assistance stops.

When dependence on technology and AI becomes a problem

We already depend on technology for communication, transport, work, and access to information. Recreating every capability ourselves would be impractical. Remembering every phone number or calculating every invoice by hand is hardly a sensible goal.

The question is which abilities you still need to exercise responsibly.

A business owner can delegate bookkeeping while retaining enough understanding to question an unexpected expense. A developer can use generated code while checking its behaviour. A writer can request editing suggestions while deciding what the argument should be.

Concern grows when you cannot evaluate the work you accept, or when assistance prevents you from learning something you actually need to know. That distinction also matters for accessibility: relying on an assistive tool is not, by itself, evidence of harmful dependence.

What the research actually shows

Trust in AI can accompany less critical thinking

A 2025 study by Microsoft Research and collaborators surveyed 319 knowledge workers. Greater confidence in generative AI was associated with less reported critical thinking. Greater confidence in their own abilities was associated with more.

The study also found that critical thinking shifted toward checking information, integrating responses, and overseeing the task. Those remain meaningful forms of thinking.

This was a survey of reported behaviour and effort. It did not establish that using AI causes permanent intelligence loss. Its practical relevance is narrower: confidence in a tool should not replace scrutiny of its output.

Better assisted performance does not guarantee learning

A study published in PNAS in 2025 tested AI assistance with nearly 1,000 mathematics students at one Turkish high school. Students using a basic AI interface performed better during assisted practice but worse than the control group on subsequent unassisted assessments.

A different tutor, designed with teacher-provided information and safeguards such as giving hints, largely mitigated the learning harm.

The distinction matters. Completing an exercise with help and learning to solve it independently are different outcomes. The experiment covered particular tools, mathematics, one school, and short-term outcomes. It does not prove that every AI tool harms every learner.

Everyday technology presents a similar question

GPS provides a useful comparison. A 2024 study involving 249 participants across two preregistered studies found that reported GPS dependence correlated with poorer navigation performance. It also found that people adjusted their use according to familiarity and often used GPS to supplement their knowledge.

That is an association, not proof that GPS causes broad cognitive decline. It does show why frequency alone is a weak measure of dependence: checking traffic and blindly following every turn are different behaviours.

What the video gets right—and what needs qualification

Kamath’s examples make a useful concern concrete: a tool may remove opportunities to practise a skill you want to retain.

But his experience cannot establish what caused his perceived decline. His prediction about becoming less capable over five years is speculation. The clip’s explanation connecting human evolution to increasing demands on memory and thinking is too broad to treat as established evidence.

An individual’s changing habits over a few years do not establish a biological trend across humanity. The defensible claim is narrower: some uses of technology can undermine independent performance or learning in specific settings. A universal verdict about human intelligence is unsupported.

A practical check: what can you still do yourself?

Use these exercises to identify skills you want to maintain. They are reflection prompts, not a validated test or a diagnosis. Choose a low-stakes task, then compare your attempt with the assisted result.

• Writing: Outline a short email before opening an AI tool. Can you state your purpose, supporting facts, and requested action? After editing, can you explain why you accepted each important change?
• Calculation: Estimate a routine total before using a calculator. For example, 45 × 45 is (40 + 5)²: 1,600 + 400 + 25 = 2,025. The useful habit is checking whether the result makes sense.
• Navigation: Before a familiar trip, recall the main route and a few landmarks. Compare with the map before setting off. Keep navigation available when needed; practise without creating a travel risk.
• AI-assisted work: Pick one recommendation or generated passage. Explain the reasoning in your own words and verify its key factual claims against the underlying sources.
• Technical tasks: If you accept generated code or an automation, describe its inputs, expected behaviour, and failure cases. Try it in a controlled environment and inspect the result before using it for real work.

Being slower without a tool is expected. The more useful question is whether you retain the understanding needed for the responsibilities you hold. If you cannot explain an output, identify what you need to learn before relying on it.

Habits that preserve judgment while using AI

Make an initial attempt. For skills you want to develop, write the outline, estimate the answer, or propose a solution before requesting help. This creates something concrete to compare with the tool’s suggestion.

Ask for feedback that leaves work for you. Request a hint, a counterargument, or an explanation of an error. Then solve or rewrite the problem yourself. Treat this as a practice habit, not a guarantee that any chatbot becomes an effective tutor.

Verify claims before acting. Open the cited source. Check whether it supports the statement and applies to your situation. A fluent explanation is not sufficient evidence.

Revisit important skills without assistance. Choose tasks relevant to your work or learning goals. You do not need to preserve every manual task simply because it once required effort.

Review the result, including visual output. If you use AI for images, inspect whether the asset meets the brief and is suitable for its intended use. The site’s ChatGPT image prompt collection provides examples to work from; choosing and evaluating the result remains your responsibility.

These are practical recommendations, not a tested programme for preventing cognitive decline.

So, are we becoming too dependent?

Some forms of dependence deserve concern. The evidence is strongest when we examine a particular task and measure what people can do after help is removed. It is much weaker when stretched into a claim that technology is making humanity less intelligent.

Keep using tools where they help. Decide which skills you need to retain. Check whether your workflow develops those skills or quietly removes every opportunity to exercise them.

For work involving AI automation, make human review and a usable fallback part of the process. If you need help planning that workflow, book a discussion.

Frequently asked questions

Is daily AI use a sign of overdependence?

Daily use alone does not establish harm. Consider whether you understand the output, can check important claims, and retain the skills your work or learning requires.

Does AI make people less intelligent?

The studies discussed here do not establish a general decline in intelligence. They identify narrower concerns involving reported critical thinking and independent performance in specific settings.

Should I stop using calculators, GPS, and AI?

Blanket avoidance is unnecessary. Use tools deliberately, practise skills that matter to you, and retain appropriate ways to check results and continue essential tasks when assistance is unavailable.

This article is for knowledge and informational purposes only. Test the recommendations in a staging or controlled environment and confirm they are suitable for your setup before implementing them on a live website or production system.