The Human Job in the Age of AI

I use AI every day.

It can turn a rough thought into working code, find a pattern in a mountain of information and produce a first draft before I have finished my coffee. Occasionally it also produces complete nonsense with the confidence of a consultant presenting slide 84.

Useful. Impressive. Slightly dangerous.

Then comes the part the machine cannot carry for me. Is the answer true? Is it fair? Does it solve the right problem? Who gets hurt if we are wrong?

That is the human job.

Most conversations about artificial intelligence obsess over capability. How smart will the next model be? Which profession disappears first? When does the machine become better than all of us at everything, including writing worried opinion pieces about machines?

I think we are staring at the wrong exam.

Human beings will never beat computers at speed, memory or pattern recognition. We settled that argument when a pocket calculator became better at division than almost every child on Earth. Competing harder would be a rather sad use of a civilisation.

Thriving requires deliberate redesign. Work, school, healthcare and government need to protect the things that make people useful to one another: judgment, care, creativity, physical presence and moral responsibility. Bolting AI onto a broken system simply automates the breakage.

Give the machine the paperwork

Give a nurse more time with patients, free a teacher to mentor a child through a difficult idea and let a small-business owner understand the numbers without hiring an oracle in a suit.

This is where AI earns its electricity.

Let it search, calculate, summarise and handle the routine. Humans can concentrate on the moments where context matters and somebody must take responsibility.

Many companies will get this backwards. They will use AI to remove people while keeping every stupid process intact. The remaining employees will supervise six dashboards, answer messages generated by other machines and attend a workshop about wellbeing.

Dear management, efficiency is only useful when somebody benefits from it.

The better approach starts with a human problem. Find the administrative sludge. Remove it. Measure whether customers receive better service and whether employees have more room to think. Keep a person in charge when the consequences are serious.

New jobs will appear around this arrangement. People will train systems, interpret results, audit decisions and design experiences. Care, craftsmanship, negotiation and live performance will gain value because presence becomes scarce in a synthetic world.

We should also prepare for the turbulence. Portable benefits, lifelong learning accounts and targeted income experiments deserve serious attention. A safety net can give people enough stability to retrain, start a business or care for somebody. That is economic infrastructure for contribution.

School is training for the wrong race

Education still rewards children for remembering answers and reproducing them on command. A machine can now do that in seconds, with excellent spelling and no need for a lunch break.

The old model has reached its expiry date.

Children need to learn how to ask better questions, test a confident answer, combine knowledge from different fields and make a decision when the evidence is incomplete. Emotional intelligence and ethical reasoning belong in the core curriculum. They are practical skills for a world full of machines that can recommend an action without caring about the consequence.

AI tutors can help. A child who needs another explanation at 10 o'clock at night can get one without feeling embarrassed. Lessons can move at the learner's pace. Teachers then gain time for motivation, social development and the glorious human work of noticing that a pupil has gone quiet.

This shift continues after graduation. Learning can no longer be a large parcel delivered during youth and expected to last for 50 years. It needs to become a normal part of adult life: affordable, recurring and connected to real work.

We already update our phones every few weeks. Updating our own skills once a decade seems oddly relaxed.

Presence will become a premium product

The more synthetic interaction we create, the more valuable genuine presence becomes.

You notice it at a live concert because the singer is in the room. The choices and small imperfections of its maker give a handmade table its value. Good eldercare depends on seeing a face, a hesitation or a story that has been told before and still deserves to be heard.

These activities have often been undervalued because their output is hard to scale. AI changes the calculation. When digital production becomes almost free, embodied attention becomes the scarce asset.

That should affect how we invest. Nursing, mental-health support, early childhood development and community organising are central infrastructure for an automated society. Isolation will not be cured by a friendlier chatbot. People need people. Annoying, complicated, slow people.

We also need to protect our attention. An AI designed to help you finish a task is a tool. One designed to keep you scrolling until 01:37 is an extraction machine wearing colourful buttons.

Choose carefully.

Human agency needs legal protection

Personal habits will not carry the whole burden. Powerful systems need public rules.

A person should retain the right to understand and challenge a decision about medical treatment, employment, criminal justice or public support. Human review must be real. A rubber stamp beside an algorithm is theatre.

Transparency matters because responsibility evaporates quickly inside a complicated system. Vendors blame the data, organisations point at the model and the model, rather conveniently, has no telephone number.

Somebody must own the decision.

As AI becomes more intimate, mental privacy and consent become basic rights. Our thoughts, preferences and vulnerabilities cannot become raw material simply because a device found an efficient way to collect them.

International cooperation matters too. Companies and countries will be tempted to weaken safety and labour standards in the hope of moving faster. That race ends with everyone arriving quickly at a worse place.

Democratic oversight will be slower than a product launch. Good. High-stakes power should occasionally encounter a locked door and a person asking difficult questions.

Productivity needs a purpose

AI can lower the cost of creating almost anything. A teenager can prototype a company. A scientist can search a field faster. A local group can analyse environmental data that once required a large institution.

This is the hopeful part.

More productivity gives us choices. We can use the capacity for better diagnostics, drug discovery and preventive care. Climate models can improve while conservation teams track biodiversity with tools that were previously out of reach. Artists can explore forms that were too expensive yesterday.

None of this happens automatically. Markets are excellent at finding demand, including demand for rubbish. Purpose still requires a human decision.

Local skills will help us keep our footing. Growing food, repairing objects, making things and organising a neighbourhood create resilience that cannot be downloaded from a server on another continent. Philosophy and spiritual life matter for the same reason. Values remain ours to choose; a generated answer cannot carry that responsibility.

Ambiguity will remain. So will grief, love, duty and the peculiar business of deciding what a good life looks like.

No software update is scheduled for that.

What I would change now

• Employers should automate the dull task and judge success by the time returned to people.
• Schools should teach verification, judgment and collaboration alongside the tools themselves.
• Governments should protect human decision rights and make lifelong learning economically possible.
• Individuals should keep one embodied skill, defend their attention and spend regular time with other humans in the same physical room.

These are choices we can make now. They require no artificial general intelligence, orbital data centre or breathless keynote.

AI is a multiplier. It will amplify our incentives, our institutions and our habits. A humane society can use it to widen access to knowledge, improve health and give people more time for useful work. A careless one will produce bureaucracy at astonishing speed.

The machine has kindly offered to take over repetition.

Let us take responsibility for everything that remains.