AI may indeed be a fantastic new tool, but with it comes new personal risks. One that you really do need to be aware of is termed “Cognitive Surrender”.
What is that?
I’ll come to that shortly.
Like many of you, I am finding real value in AI.
Not in some abstract futuristic sense. I mean right now, in ordinary life, at work and at home.
At work, for example, we recently ran a series of workshops over Microsoft Teams. We recorded them and got transcripts. Yes, Teams is plugged into Copilot. Yes, it can summarize meetings and extract actions. But I also found that if I dropped the full transcripts from multiple workshops into an AI tool, I could start asking much more useful questions.
For example: give me a table of every component we discussed, list the version number if known, and briefly summarize what was said about each one.
Boom! There it was.
I have experienced similar wins at home.
I recently had a go at restoring an old kitchen table and made a proper mess of it. Before long, the paint was coming off and I was wondering how on earth to rescue it. So I described exactly what I had done, and asked ChatGPT how to fix it. It gave me a step by step plan, told me what materials I needed, and explained why the first attempt had failed.
I followed the advice and, somewhat to my surprise, it really worked. (Phew!)
Recently, just a few days ago, I ran into a very different example. I came across a Medium article titled AI Finds Jesus’ Resurrection Best Explains the Historically Verifiable Facts. The author’s claim was that AI is brutally honest and boldly points to the Resurrection.
My immediate reaction was that this smelled strongly of BS, but did I really want to spend time wading through the whole thing to work out what the problem was?
Then I remembered that AI can also be useful for analysis.
So I dropped the article into ChatGPT along with this simple prompt: Please review the following claim and advise on what the most honest fact-based position really is.
The answer, in summary, was this (you can also find the full answer posted as a comment under the article):
The resurrection is a live theological interpretation of the evidence, not a historically proven conclusion, and the article exaggerates both scholarly consensus and what AI can do.
Bingo; a clear honest answer that did not vent and was not sneering. It was simply honest and spelled out how it really was.
And that, I think, gets us to the heart of what AI is actually good for, and is now being actively used to do because it does it well.
What is AI really being used for?
We do not have to guess. There is plenty of data on this, and so the broad picture is clear.
In practice, AI use seems to fall into a few big buckets.
First, there is writing and office work. People use AI to draft emails, summarize meetings, rewrite documents, extract actions from notes, answer questions over files, and help with coding.
Second, there is search and question answering. A lot of consumer use is simply wanting a quick explanation, a first draft, or a fast answer to a practical question. I guess you can think of it as Google on steroid perhaps.
Third, there is image, audio, and media work. AI is now routinely used to clean up images, remove backgrounds, generate visuals, improve audio, create subtitles, and produce marketing material.
Fourth, there is customer service and internal support. Companies use AI to draft replies, route tickets, power help centres, and assist staff dealing with routine queries.
Fifth, there is education and tutoring. Students use it for explanations, revision help, practice questions, language support, and homework assistance.
Sixth, there is specialist work such as software development, fraud detection, forecasting, healthcare administration, document review, and data analysis.
That last point matters, because in most of these cases the real pattern is not full automation. It is assistance.
AI is not usually replacing the whole job. It is taking the friction out of parts of it.
That is why AI already feels common. It is not because everyone has handed their life over to a chatbot. It is because useful tools spread fast when they save time, remove hassle, and widen access to skills people did not previously have.
Used well, AI is genuinely impressive.
Used badly, it can make us stupider.
Hang in there, I’m getting to the point here, it comes within the next two sections.
Cognitive offloading is normal
A good way to think about AI and how it is impacting our lives is through the idea of cognitive offloading.
That sounds technical, but the idea is simple. It means using something outside your own head to reduce mental effort. Humans have always done this. We write lists. We use calendars. We set reminders. We use calculators. We follow satnav. We keep notes. We build systems around ourselves so we do not have to hold everything in our heads at once.
That is normal.
I can still remember the landline number we had when I was growing up. But I do not know any of the phone numbers I use now. My phone knows them. When I drive somewhere unfamiliar, I use GPS. I no longer unfold a giant paper map and work it all out in advance. When I need to remember something important, I set a reminder.
None of that is weird. It is what we all do.
The important point is that with cognitive offloading, you are still in charge. Something is helping you store, track, retrieve, or process something. It is supporting your thinking, not replacing your judgment.
That can be smart.
But there is another step beyond that.
Cognitive surrender is where the danger begins
The real risk is not simply using AI. It is what happens when you stop really evaluating what AI gives you.
This is where the idea of cognitive surrender comes in.
The basic idea is straightforward. Offloading is letting a tool help with part of the task. Surrender is giving up critical oversight and simply going with whatever the tool says because it is fluent, fast, and sounds authoritative.
Think of it like this:
- Cognitive offloading is using GPS to get from A to B.
- Cognitive surrender is obeying the GPS so blindly that when it tells you to turn left into a pond, you do it.
That is obviously absurd in that form, but the same pattern can happen with AI. The wording is polished. The tone is calm. The answer arrives instantly. It sounds as if it knows what it is talking about. So people stop checking.
That is where things can go badly wrong.
You ask AI to summarize a long report, then compare the summary to the report and decide what you think. That is cognitive offloading.
You ask AI what the report means, never read the source, and then repeat its answer as if it must be correct. That is cognitive surrender.
Back to that resurrection example.
For someone already invested in the belief, “AI says it” can become a shortcut to certainty. The machine has spoken. Case closed. No further thought required.
That is not evidence of truth. It is evidence of misplaced trust.
And this can become far more serious than one silly religious article.
Many people are now using AI for health information, technical advice, financial questions, legal explanations, and all kinds of other things where bad answers can have real consequences. The problem is not that AI is always wrong. It often is useful. The problem is that it can also produce total BS in a tone that sounds polished and confident.
Once people start treating fluency as truth, they are in trouble.
I wrote about this recently in the context of medical advice:
Offload the burden, not the responsibility
This is the key:
Offload the burden, not the responsibility.
Use AI for memory, speed, organization, drafting, summarizing, and idea generation. Use it to help you get started, compare options, interrogate a messy transcript, or make sense of a chaotic first pass.
But keep the judgment for yourself.
Check the source. Read the document. Verify the quote. Compare the answer against reality. Ask whether it actually addresses the question. Notice what has been left out.
AI can be a very good assistant.
It is a very bad thing to treat as an oracle.
One last caution
There is one other lesson I have learned.
AI is only as good as the questions you ask it.
If your prompt is vague, key details may be missing. If you leave out an important constraint, it may confidently give you an answer that is technically polished but practically useless.
I have definitely seen that happen.
It reminds me of the following Pink Panther clip where a question was asked, “Does your dog bite?“. The answer, while accurate, was incomplete:
That is often how AI works as well.
Ask a lazy question and you may get an accurate but incomplete answer.
Ask a far more precise question and things improve dramatically.
So where does that leave us?
For me, AI is already clearly useful.
It helps me at work. It helps me at home. It helps me summarize and analyse large amounts of text, extract structure from chaos, and get faster at tasks that used to take much longer.
But I also think the real danger is not that AI will become some all-powerful mind that takes over everything.
The real danger is more ordinary than that.
It is that people start handing over too much of their own judgment because the machine sounds convincing.
That is the risk.
Use it, by all means.
Don’t let it do your thinking for you.
Questions for Commenters
- Do you use AI, and if so which one, what for, and how useful was it?
- Do you get real value from it?
- What tips have you picked up?
- And have you found anything genuinely cool?
References
Pew Research Center. (2026, March 12). Key findings about how Americans view artificial intelligence.
Singla, A., Madgavkar, A., et al. (2025, March 12). The state of AI: How organizations are rewiring to capture value. McKinsey & Company.
Stanford Institute for Human-Centered Artificial Intelligence. (2025). Artificial Intelligence Index Report 2025.
Risko, E. F., & Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences, 20(9), 676–688.
Shaw, S. D., & Nave, G. (2026). Thinking—Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender. SSRN / PsyArXiv.
Pew Research Center. (2025, February 25). Workers’ experience with AI chatbots in their jobs.
Pew Research Center. (2025, February 25). U.S. workers are more worried than hopeful about future AI use in the workplace.
McClain, C., Anderson, M., Sidoti, O., & Bishop, W. (2026, February 24). How Teens Use and View AI. Pew Research Center.
Associated Press. (2026, April 15). Why many Americans are turning to AI for health advice, according to recent polls.