The AI Spiral: Why the Next Evolution of AI Literacy Is Knowing When Not to Use AI

Imagine driving a large SUV just a few hundred metres to buy a loaf of bread.

Could you have walked? Probably.

Would it have taken much longer? Probably not.

Will that one trip change the planet? Of course not.

But imagine millions of people making that same choice, several times a day.

Small decisions, repeated at scale, have consequences.

I wonder if we’re beginning to treat AI the same way.

Not because AI isn’t useful — it absolutely is. AI is transforming healthcare, accelerating scientific discovery, making knowledge more accessible, and helping people solve problems that were once out of reach.

But we’re also increasingly reaching for one of the world’s most powerful computing systems to do things we could often do ourselves, just as quickly.

The question isn’t whether AI can do these things.

It’s whether we should ask it to.

The AI Spiral

A few days ago, I caught myself doing exactly that.

I was writing a simple email that would have taken me two minutes. Almost instinctively, I opened my AI assistant.

“Make it warmer.”

“A little shorter.”

“More professional.”

“More like me.”

A few prompts later, I paused.

I’d spent more time refining the email with AI than it would have taken to write it myself.

That made me wonder:

Had AI become my first instinct instead of my best tool?

I call this The AI Spiral — when the ease of prompting leads us into repeatedly refining work we were already capable of doing ourselves.

The AI Spiral usually follows three stages:

• Instinct — We reach for AI before considering whether we actually need it.
• Iteration — Another prompt feels almost effortless, so we keep refining.
• Illusion — We believe we’re being more productive, even when we’ve spent longer than doing the work ourselves.

The technology isn’t the problem.

The habit might be.

Convenience Isn’t Always Productivity

Humans naturally gravitate towards the path of least resistance. AI has reduced the effort required to generate content to almost zero.

That’s one of its greatest strengths. But it also creates an unintended consequence.

We may begin outsourcing decisions that were never burdensome to begin with.

We’ve all become familiar with using AI to:

• Rewrite emails.
• Draft meeting invitations.
• Rephrase Teams messages.
• Adjust the tone of a paragraph.
• Generate images or videos for fun.
• Chat out of curiosity or entertainment.

and many more simpler activities, easy and swift for human judgment

None of these activities is inherently wrong. Many create genuine value through creativity, accessibility, learning, and communication.

But somewhere along the way, convenience can quietly become habit. And habit can gradually replace judgment.

One Prompt Doesn’t Matter. Billions Do.

This isn’t about one email.

It’s about scale.

Multiply one unnecessary five-minute AI interaction by 5,000 employees, several times a day.

Beyond the computing resources consumed, think about the collective loss of time, attention, and independent judgment.

Now multiply that across millions of organisations and hundreds of millions of users.

AI assistants now support hundreds of millions of people worldwide, generating billions of interactions. At the same time, the International Energy Agency (IEA) projects that electricity demand from data centres is expected to more than double between 2024 and 2030, with AI being one of the major drivers.

I’m not suggesting that rewriting one email harms the environment. Just as one unnecessary SUV trip doesn’t.

But billions of low-value interactions, repeated every day, contribute to growing demand for computing infrastructure, electricity, and cooling.

The environmental impact isn’t created by one prompt. It’s created by our collective habits.

The Next Evolution of AI Literacy

Reflecting on The AI Spiral made me rethink what AI literacy should really mean.

Today, AI literacy is often measured by how well we use AI — writing better prompts, exploring new models, understanding bias, transparency, accountability, and hallucinations, and producing better outputs.

These are foundational skills and will continue to be essential.

But I believe the next evolution of AI literacy is different.

It’s not just about knowing how to use AI.

It’s about knowing when to use it.

I call this AI Judgment.

When we’re young, many of us are taught simple habits that become second nature: switch off the lights when you leave a room, turn off the tap while brushing your teeth, don’t waste paper, recycle where you can.

None of those individual actions saves the planet on its own.

We teach them because they build a mindset — that shared resources deserve thoughtful use and that small actions, repeated by millions of people, make a meaningful difference.

Perhaps it’s time to extend that same mindset into the AI era.

Today, a prompt feels almost free. It’s instantaneous, effortless, and always available.

But while it may feel free to us, it isn’t resource-free. Every prompt relies on computing infrastructure, electricity, and cooling.

Individually, the impact of a single interaction is small.

Collectively, billions of interactions shape the environmental footprint of AI.

Whether we’re rewriting an email, generating images for fun, creating a video we’ll watch once, or chatting with AI out of curiosity, the question isn’t whether these activities are right or wrong.

The question is whether they’re intentional.

Just as digital literacy taught us to think about privacy, security, and our digital footprint, AI literacy should also encourage us to think about our AI Footprint.

The future won’t belong to the people who use AI for everything. It will belong to those who know where AI creates meaningful value and where their own judgment is already enough.

AI Should Replace Effort, Not Thought

AI is at its best when it helps us do what would otherwise be difficult or impossible.

Analysing vast amounts of data.

Accelerating scientific research.

Making information more accessible.

Supporting better decisions.

These are the kinds of problems AI was built to solve.

But asking AI to rewrite the same two-line email five times?

That isn’t always productivity.

Sometimes it’s simply a habit.

Writing, editing, and choosing the right words are how we practise judgment, communication, and critical thinking skills that become even more valuable in an AI-enabled world.

The biggest risk isn’t that AI thinks for us.

It’s that we gradually stop noticing when we’re no longer thinking for ourselves.

Responsible AI Starts With Us

Responsible AI is often discussed in terms of governance, fairness, security, and compliance.

Those conversations are essential. But responsibility also lives in everyday choices.

People are the zeroth line of defence, and AI is a tool. How we choose to use that tool ultimately determines both its value and its impact.

Before opening an AI assistant, perhaps we should ask ourselves:

• Is AI genuinely adding value here?
• Would I complete this faster myself?
• Am I using AI because I need it — or simply because it’s available?
• Am I solving a real problem, or optimising something already good enough?

These aren’t just productivity questions. They’re AI Judgment questions.

And increasingly, they’ll become leadership questions.

The Better Question

We don’t ban cars because walking exists. We simply choose the right mode for the journey.

AI deserves the same mindset. Not every task needs the most powerful tool available.

Before your next prompt, pause for a moment and ask yourself:

If AI didn’t exist, would I have completed this task faster myself?

If the answer is yes, perhaps the most responsible decision isn’t writing a better prompt. It’s relying on your own judgment.

Because the future of Responsible AI won’t be defined only by the systems we build.

It will also be shaped by the habits we develop.

Responsible AI isn’t just about governing intelligent systems. It’s about governing our growing dependence on them.

References

• International Energy Agency (IEA). Energy and AI: https://www.iea.org/topics/artificial-intelligence
• International Energy Agency (IEA). Electricity 2025: https://www.iea.org/reports/electricity-2025
• OECD AI Principles: https://oecd.ai/en/ai-principles
• UNESCO Recommendation on the Ethics of Artificial Intelligence: https://www.unesco.org/en/artificial-intelligence/recommendation-ethics