We're Not Predicting the Future of Work.
We're Choosing It.

I'm Rob Devlin, Founder and Managing Partner of Devlin Talent, a global talent acquisition consultancy that helps businesses achieve strategic and digital transformation.

I'm deeply honored to be leading the Case Study workstream of the Human-Centric AI Council throughout 2026 and beyond. Our workstream is dedicated to capturing case studies from organizations that are implementing, or have already implemented, AI solutions. Our role is to document and analyze their experiences, build greater awareness of the important work happening in this space, and help other organizations navigate this rapidly emerging area of work.

As part of this leadership role, my colleagues and I have been thinking about the speed of AI development, and all the confident predictions about what work will look like in five or ten years.

And of all the days to write this blog, I've chosen August 29, which AI enthusiasts will all know is the day that Skynet becomes self-aware at 2:14 a.m. EDT.

As it stands, it's 12:21 p.m. PST, so I think we are going to be okay. But maybe a Judgment Day theme for this blog could be an interesting angle.

What I'm noticing more and more is that even short-term predictions around AI can quickly become outdated. Predicting individual tools or the capabilities of AI tools is definitely NOT the same as understanding their effects on work.

I've had some recent firsthand experience of this while going through a round of AI vendor assessments (pretty rapidly, I might add) for a big tech firm I am consulting for. Even by the end of the assessment cycle, some of the products had evolved significantly.

So, because of this, I am thinking that the most useful question we should be asking right now is:

---

Forecasting Judgment Day versus choosing the future

Pop culture would have us believe the future of AI is an inevitable destination: technology develops, control is lost, and we'd all better call Arnold because Judgment Day is not far behind.

However, I'm realizing more and more that real organizational futures are created through many human decisions, not one dramatic moment when a machine becomes conscious. The interesting thing here, though, is how conscious are WE that we are making those human decisions right now, and that those decisions are the ones shaping our future?

Same subject, three questions

Forecasting

What is likely to happen?

Futures thinking

What could happen?

Human-centric futures thinking

What should happen, and what choices could take us there?

I don't think, in our day-to-day work, we are trying to produce one perfect prediction of the future. We are imagining multiple possible futures and deciding which ones are worth building.

---

We might recognize Skynet.
But do we know how it will change the future of work?

We can already see that AI is becoming more autonomous and capable of acting across more systems. At the same time, we are all eager, and probably guilty, of feeding it more and more data to see it do more and more cool stuff. Its power and pace are seductive, for sure.

Then we have agents, MCPs and headless systems entering the chat, making the technology less visible, more influential and increasingly difficult to keep track of. Very mysterious. Very powerful. Potentially very Skynet.

This leads us down a path of increasingly blurred lines and boundaries. What exactly IS AI doing here? Where does one system end and another begin? And, perhaps most importantly, who remains accountable for the decisions being made across them?

This is also where some established roles may disappear, combine or become something entirely different. As with any industrial revolution, the boundaries between customers, workers, contractors and people whose data is being used could become increasingly difficult to distinguish. I think many of us will occupy several of these overlapping roles, sometimes without even realizing it.

Because of all this, we need to stay super focused on the fact that, while the direction of technological development may be increasingly visible, its organizational and human consequences are not.

---

There is no single red button

This is where things get interesting. The idea that there will be one big red button, one dramatic decision to activate or deactivate Skynet, is not how this is going to happen. It's ALREADY happening through hundreds of decisions that we are all having to make on the daily, many of which may initially appear operational or tactical.

Those decisions are increasingly being made, consciously or subconsciously, through prompt engineering, AI design and the way we choose to implement these systems.

Where conscious thinking is taking place, we need to stay focused on one very important question.

Do we use AI to replace people, augment them or redesign the work?

ReplaceThe work stays; the people doing it don't.

AugmentThe people stay; the work gets easier.

RedesignThe work itself changes shape.

Then we need to ask

1. Are the people affected involved in the decision?
2. Can those people understand, question or challenge an AI-supported outcome?
3. Who exactly benefits from the productivity created?
4. What happens to the capacity AI releases?
5. Who remains accountable when decisions move across interconnected systems?

---

Who actually gets to live in the future?

In Terminator 2, it's pretty clear who is building the machines, who is trying to survive them and who gets to make a dramatic entrance in head-to-toe leather. Back in the real world, though, those lines (and the dress code) are becoming far less obvious.

Let's start with the employee. All of a sudden, this feels way too narrow and too layered for the future we're describing.

AI systems affect far more people than this. They affect managers, employees, candidates, contractors, gig workers, customers and people whose data is being used to train or operate them.

And don't forget, people can occupy several of these roles at any one time. Also, let's be real: SOME people are affected by workplace AI without ever choosing to participate in it.

Candidates are a good example, because AI can, and does, make consequential decisions about them before they ever become employees.

So, who has agency? Who understands the system, and who can challenge it?

For me, it's massively important that the future works for more than just the leaders and technology companies designing and profiting from it.

---

The case studies:
evidence sent back from the future

Unfortunately, we don't have a hunky, reprogrammed T-800 to send back and tell us exactly where our AI decisions are going to lead. What we DO have are real organizations making those decisions right now, and their experiences can give us some pretty important clues.

And this is where I need your help. As part of my work with the fabulous HCAIC, I am working alongside an amazing workstream team to gather case studies from real businesses and real situations, so that we can ground these possible futures in real organizational experiences.

We aren't only looking for polished AI success stories, although these will be very welcome.

We want the good, the bad and the "someone should probably have seen that coming."

The case studies we are looking for should capture the original problem, the choices made, who was affected and what really happened. They should tell us who gained or lost agency, visibility or power. Failures, unintended consequences and lessons learned are, in my opinion, just as useful, if not more useful, than perfect outcomes.

We're not trying to predict the future from individual examples. We're going to be identifying recurring signals and patterns, and the decisions and consequences already shaping it.

---

Choose your own Judgment Day

Gathering all this evidence into a beautifully formatted report that lives untouched in a mystery SharePoint folder would be its own kind of Judgment Day. Instead, what if we could turn these real-world decisions into a high-level "choose your own adventure" for the future of work?

Users could enter a specific business challenge and make a series of AI implementation choices. Each decision would then nudge them towards a different possible future, while showing the impact on the different people participating in the system.

Of course, there isn't one perfect answer. The purpose is to reveal the trade-offs. Users could then work backwards from the future they want to identify the better decisions they need to make today.

This could start as a simple, facilitated decision-making toolkit and perhaps eventually become an interactive app. But the important thing is to get the logic, choices and possible consequences right before we get distracted by building something shiny.

It wouldn't claim to predict what WILL happen. It would give us a safe space to explore what COULD happen, understand how today's decisions might take us there and decide whether we really want to follow that particular timeline.

---

No fate but what we make

Nobody can tell us exactly what the future of work will look like. Not even someone sent back from the future with suspiciously detailed knowledge and an excellent leather jacket.

But uncertainty doesn't remove our responsibility to make conscious decisions now. Every time an organization chooses where, why and how to deploy AI, it makes one possible future a little more likely.

Human-centric AI isn't simply about making technology safer, faster or more efficient. It's about deciding what kind of relationship we want to create between people, technology and work, and making sure that the people who will live with the consequences have a voice in shaping it.

As Sarah Connor reminds us: "There's no fate but what we make."

So, if your organization has implemented AI at work, or even attempted to, and you have a story about what really happened, we would love to hear from you. We want the successes, the failures, the unintended consequences and the moments when someone really should have seen it coming.

Together, we may not be able to predict the future of work, but we can become much more deliberate about the one we are helping to create.

Now, I'm off to put on my dusty Sarah Connor wig and apply my factor two-million sunblock.

JK. I'm off to feel good about all the fabulous futures we can responsibly build with AI.