Should humans have to adapt to AI?

The jobs people will be doing five years from now may not exist today. That uncertainty is already influencing education, careers and organisational decisions. It also raises a deeper question about the relationship humans are building with technology. As AI capabilities advance, people are increasingly being asked to change how they work, learn, communicate and even think about their own value.

In Episode 18 of The AI Values Podcast, Lambert Hogenhout, Chief of Data and AI at the United Nations Secretariat, joins Edosa Odaro and Lindley Gooden to examine what AI means for work, education, organisational culture and human happiness. His argument pushes leaders to consider the kind of progress they are trying to create and the human qualities they want to preserve along the way.

As Hogenhout puts it:

“I'm very sad that we ask ourselves this question because really, technology should be adjusting to us and we shouldn't have to reinvent ourselves to technology.”

That idea runs through the entire conversation.

AI will continue to develop. Leaders still have choices about how organisations respond to it, where it is deployed and which parts of human experience deserve protection.

The jobs five years from now may not exist yet

Predictions about the future of work often begin with today's jobs.

Which roles will disappear? Which tasks will become automated? Which careers will remain safe?

Hogenhout sees a problem with trying to forecast the future through today's organisational structures.

When he speaks with university students, he encounters understandable anxiety. Young people want to know which careers will remain viable and what they should study to prepare themselves.

Precise answers are difficult because new technologies also create forms of work that previously had no name.

“The jobs five years from now don't exist yet.”

Hogenhout points to the emergence of the internet. Careers associated with search engines, social media and digital platforms developed alongside technologies that created entirely new forms of economic activity.

AI could produce a similar shift.

Some existing tasks will change. Some roles may disappear. New ones will emerge. Organisations will redesign work around capabilities that previously did not exist.

That uncertainty changes the leadership challenge.

Workforce planning needs to give greater weight to adaptability, curiosity, judgement and the capacity to learn. Education faces the same pressure.

Preparing people for a single predictable career path becomes increasingly difficult when the underlying technology keeps changing.

The ability to learn may become one of the most durable capabilities organisations can develop.

Efficiency is too small a goal

Organisations naturally want measurable returns from technology.

AI can make processes faster. It can automate repetitive work. It can increase output and potentially reduce costs.

Hogenhout argues for a wider organisational view.

“Organizations need to think about this holistically and not just in terms of efficiencies.”

An efficiency measure captures only part of what happens when technology enters an organisation.

A deployment can improve productivity while changing morale. It can reduce costs while affecting trust. It can accelerate communication while making interactions feel less authentic.

These consequences can influence organisational performance long after the original efficiency gain has been recorded.

Hogenhout therefore encourages leaders to stay close to organisational identity.

“Stay true to who you are as a company and what is your culture, what are your values?”

That is a governance question.

AI decisions shape how work is experienced. They affect relationships between employees and employers. They can alter customer interactions and change where responsibility sits.

Boards need to understand those effects alongside the financial case.

We may be automating the wrong things

One of Hogenhout's most memorable observations concerns the activities AI is currently becoming good at.

“We have AI that does all the things that I like to do, like writing poetry or fun stuff. And I'm still taking out the trash by myself.”

There is humour in the line, but the underlying question matters.

What should automation remove from human life?

Organisations frequently approach automation through cost. Expensive activities attract attention because removing or reducing them appears to offer an obvious financial return.

Hogenhout suggests another starting point: ask people about the work they would genuinely like technology to take away.

“What are the things that you least like about your job? What are you hate doing?”

The grammar reflects the spoken transcript, but the point is clear.

Employees already know which tasks consume time without providing much meaning. Administrative repetition, unnecessary processing and frustrating routine work may provide useful starting points for automation.

That approach could also improve adoption.

People may respond differently to AI when they experience it removing friction from their working lives rather than threatening activities they value.

For leaders, this means employee experience deserves a place in decisions about automation.

AI can create value and destroy it simultaneously

Efficiency gains can hide consequences elsewhere in the organisation.

Hogenhout describes an example in which AI was introduced into a process and parts of the work were automated. The expected productivity gains appeared attractive.

The human consequences were less positive.

“They got all demoralized. They didn't trust the input that they got from AI or if they had to hand over their work to AI, they didn't get any satisfaction out of that.”

This is an important warning for leaders measuring AI value.

A process can become more efficient while employees become less engaged. Automation can increase output while weakening confidence in the work. A local improvement can create costs elsewhere.

Those effects may eventually appear through retention, morale, quality, customer experience or trust.

AI value therefore needs to be measured across a longer time horizon.

Financial return matters. So do organisational culture, reputation, employee experience and customer confidence.

The broader leadership question becomes: what outcome are we trying to create, and is this technology helping us reach it without sacrificing something more valuable along the way?

Happiness belongs in the AI conversation

Hogenhout brings an unusual measure into the discussion of responsible AI: happiness.

“The ultimate goal in life, I think, for many of us is to be happy.”

The idea can sound distant from boardroom discussions about productivity and AI investment.

Yet organisations already make decisions that influence people's experience of work. Technology affects autonomy, relationships, satisfaction, purpose and the sense that an individual's contribution matters.

Those factors have organisational consequences.

If AI removes frustrating work, employees may gain more time for meaningful activity. If it removes work that provides identity or satisfaction, the effect may run in the opposite direction.

Hogenhout's discussion of happiness broadens the definition of value.

Leaders can ask whether technology helps people do work they consider worthwhile. They can examine whether AI strengthens or weakens relationships. They can consider what employees gain from the time technology saves.

This creates a more demanding measure of AI success.

Productivity remains visible. Human experience becomes visible too.

AI is moving faster than society can adjust

Technological change has always required social adaptation.

AI compresses the time available for that adjustment.

Hogenhout describes the difference starkly:

“These days, we have 15 hours before something new comes out and it's not enough.”

Organisations are experiencing the same pressure.

New models, capabilities and products arrive before many organisations have fully understood the previous generation. Employees encounter tools before policies have been written. Leaders make investment decisions while the underlying technology continues to change.

Hogenhout's response is deliberate adoption.

“Slow down and be very deliberate about how you adopt technology as a human and also as a company.”

Slowing down in this context means creating enough space to make a considered decision.

What problem does the technology solve?

What behaviour might it change?

What happens to employees?

What happens to customers?

What new dependencies are created?

What should remain human?

These questions help organisations distinguish useful adoption from adoption driven primarily by technological momentum.

Governance should enable good outcomes

AI governance is often associated with controls, risk registers and restrictions.

Hogenhout gives it a broader purpose.

“Governance is on the one hand risk mitigation.”

He then adds:

“The other part of governance is you want to make sure that the right things go in the right direction.”

This creates a useful framework for boards.

Good governance protects organisations from unacceptable consequences while creating conditions in which worthwhile applications can succeed.

That requires judgement.

Every decision to deploy AI creates risks. Decisions to delay or reject useful technology can create risks as well. Organisations therefore need ways to evaluate the consequences on both sides.

Values help provide that basis.

A clear understanding of organisational purpose, responsibilities and acceptable trade-offs gives leaders something against which technological opportunities can be assessed.

Without that clarity, governance can become reactive.

With it, governance can help organisations make deliberate choices about where AI belongs.

‍

Listen to Episode 18

Going deeper: AI strategy needs a definition of human progress

Episode 18 exposes a weakness in many organisational AI strategies.

They contain measures of productivity, cost, adoption and technical performance.

Measures of human progress are harder to find.

That omission becomes increasingly important as AI reaches further into work.

If employees become more productive while engagement falls, leaders need to understand the full organisational effect.

If customers receive faster service while trust declines, the value calculation changes.

If AI increases content production while authentic human communication becomes harder to identify, output alone provides an incomplete measure of progress.

Technology metrics cannot resolve these questions.

Boards need a broader definition of value connecting financial outcomes with trust, culture, employee experience, customer outcomes and institutional purpose.

Automation decisions should reflect that definition.

Leaders can begin by identifying activities people find repetitive, frustrating or unnecessarily administrative. They can examine where technology could create more space for judgement, relationships, creativity and meaningful work.

Governance should follow the same principle.

Risk mitigation remains necessary. Organisations also need governance that helps worthwhile applications develop responsibly. Pursuing technical possibilities without understanding their human consequences can gradually weaken the conditions on which long-term organisational success depends.

Leadership therefore needs a clear view of what the organisation wants to preserve while it changes.

That leads to a question technology cannot answer:

What should progress actually mean for the people affected by it?