What Should We Let AI Do for Us - and What Should We Never Give Away?

I have been thinking a lot about artificial intelligence lately. Not from the perspective of whether it is good or bad, whether it is going to take everyone's jobs, or whether we should be excited or terrified by it. I am much more interested in a practical question: what should we let AI do for us, and what are the things we should be very careful not to give away?

Perhaps unsurprisingly, I asked AI about it.

I use AI regularly now. I use it to research, organise information, analyse things, explore ideas and help me make sense of the approximately seventeen thoughts that can be competing for attention in my head at any one time. I talk things through, question things, disagree with it, change direction and sometimes discover that the thing I thought I was talking about was not really the interesting bit at all.

This article is an example of that relationship. The thinking behind it has developed through my own work, experiences and many conversations about coaching, behaviour change, weight loss and, increasingly, AI itself. I talked through what I wanted to say and used AI to help me bring those thoughts together into a coherent article. Then I read it and asked myself a very important question:

Is this true for me?

That feels like a much more useful test than asking whether I personally typed every word.

Let AI make life easier

There are some things AI can do extraordinarily well.

Imagine opening the fridge and saying, "This is what I have. I need meals for the next four days, I am out on Tuesday evening, Wednesday needs to be something quick and I do not want to go shopping until Thursday."

Within seconds, you can have some ideas.

It can take those meals and make a shopping list. It can help you plan a morning routine, organise a complicated week, compare information, create a packing list, break an overwhelming job into manageable pieces or help you work out where on earth to start.

Used well, AI can remove friction.

That matters because friction takes energy. Every tiny decision we make uses a little bit of our capacity. What shall we eat? When am I going to do that? What do I need to remember? Where did I put that information? Which thing needs doing first?

Sometimes we do not need another hour in the day. We need fewer things competing for the hours we already have.

If technology can take some of that load, I am very happy to let it.

I have started asking myself a question about my own work:

Did this require me?

Did it require my judgement, experience, creativity or relationship with another person? Or was I about to spend two hours formatting, sorting, calculating or organising something that technology could help me do in twenty minutes?

There is no prize for doing something the hard way simply because we can.

Information has never been the whole answer

This is where it becomes particularly interesting to me, because I spend a lot of my working life around behaviour change.

I have worked with people wanting to change their relationship with food and weight for many years. I also work as a coach with people exploring changes in completely different areas of their lives.

In both worlds, I see the same thing repeatedly.

Knowing something does not necessarily mean we do it.

Most of us already know that eating differently can affect our weight. We know movement is good for us. We know sleep matters. We know constantly saying yes to everybody else while leaving ourselves at the bottom of the list probably is not helping us.

Information is important, but information alone does not always create change.

If it did, none of us would ever struggle with anything once we knew the answer.

AI is going to make information even easier to access. It can give us recipes, routines, explanations, ideas, strategies and step-by-step plans in seconds.

That is enormously useful.

It still leaves one rather important question:

Will we actually do any of it?

Sometimes we do not need another answer

There are times when the most valuable thing another person can offer us is not an answer at all.

It might be a question.

"What is really getting in the way?"

"What would you do if that was not a constraint?"

"You have said you really want this, so what keeps making something else more important?"

"What are you afraid might happen if you actually do it?"

Sometimes it is not even the question itself. It is the fact that somebody knows us well enough to know which question to ask.

There is something powerful about another human being noticing the hesitation before we answer, recognising that our energy has changed, remembering something we said three weeks ago or gently pointing out that what we are saying today does not quite match what we said we wanted.

A good conversation can take us somewhere we did not know we needed to go.

That is one of the things I love about coaching. It is also one of the things I love about working with people in groups.

The energy of other people matters.

Being noticed matters.

Being encouraged matters.

Laughing matters.

Being challenged at the right moment matters.

Sometimes simply knowing that somebody will ask us next week how we got on makes a difference.

AI can provide an astonishing amount of support, and I suspect that support will become increasingly sophisticated. In my experience, there is something fundamentally different about one human being genuinely present with another.

Perhaps it does not need to be AI or human

This is where I believe it is possible to get ourselves tied in knots.

We keep talking about AI as though we have to choose.

Technology or people.

Efficiency or connection.

AI or human beings.

I do not think that is the most interesting choice.

The opportunity is in getting much better at recognising what each is good at.

Let AI help with the meal planning.

Let it turn the plan into the shopping list.

Let it help organise the week.

Let it analyse the information.

Let it find the pattern.

Let it take the enormous thing that has been sitting on your to-do list for three weeks and help you break it into five manageable steps.

Then use the time and headspace you have gained for the things that matter.

Have the conversation.

Go to the group.

Ask for help.

Listen properly.

Spend time with somebody.

Explore why the thing you keep saying you want still is not happening.

Have the uncomfortable conversation you have been avoiding.

Ask yourself a better question.

Maybe the future is not about asking whether AI can replace human beings.

Maybe it is about letting technology do more of the things that do not require us, so we have more capacity for the things that do.

There is another side to this

There is a temptation with any technology that makes something easier to hand over more and more.

This is where I place real value on retaining our curiosity.

Just because AI can do something does not necessarily mean I want it to.

I do not want it deciding what I believe.

I do not want it replacing my curiosity.

I do not want it creating a polished version of me that sounds impressive but no longer sounds like me.

I particularly do not want to outsource the experiences, conversations, mistakes, observations and random thoughts that eventually become my ideas.

This matters to me with social media and writing.

I can ask AI to produce fifty motivational posts before breakfast. I do not want fifty motivational posts before breakfast.

I would much rather share one thought that came from a conversation I had, something I noticed, something that challenged me, something that made me laugh or something that I am genuinely still exploring.

AI can help me develop that thought. It can ask me questions about it. It can help me find the interesting part, organise it, shorten it or turn one long piece into something useful elsewhere.

The starting point still needs to mean something to me.

Perhaps that is another useful question:

Am I using AI to amplify my thinking, or am I asking it to do my thinking for me?

I am very comfortable with the first.

I am much less interested in the second.

Human does not have to mean inefficient

I wonder whether some of our discomfort with AI comes from the belief that using it somehow makes what we produce less authentic.

I am starting to see it differently.

If I spend three hours wrestling information into a spreadsheet, I have not necessarily added three hours of Sarah to it.

I have just spent three hours wrestling with a spreadsheet.

If technology can do much of that processing and leave me with the part that requires interpretation, judgement and a decision, I have not made my contribution less human.

I have concentrated the human bit.

The same applies to coaching and supporting change.

If technology can help somebody plan their week, find meal ideas, organise their thoughts or prepare for a conversation, perhaps the time we spend together does not need to be used doing those things.

We can spend it exploring what is underneath them.

That feels less like replacing human interaction and more like making space for better human interaction.

The question I am taking forward

AI is going to change how we work. It is already changing how I work.

I do not pretend to know exactly where it will take us, and I am wary of anybody who claims they do.

What I do know is that I do not want to reject something enormously useful simply because it can do things I used to do myself. Neither do I want efficiency to become the goal at the expense of everything that makes our lives human.

So I am going to keep asking:

Did this require me?

If the answer is no, perhaps there is a quicker, easier or smarter way.

If the answer is yes because it needs my judgement, my experience, my creativity, my energy, my relationship with somebody or simply my presence, that is probably exactly where I should be spending more of my time.

Perhaps that is the opportunity AI gives us.

Not to become less human.

To make more room for the parts of us that are.