AI won't decide what it does to your workforce. Your deployment choices will.

You keep being told that AI will "transform the workforce", and the phrasing does something quietly unhelpful: it makes the outcome sound settled. Some inevitable wave is coming that will either lift your people or replace them, and your job is to read the forecast and brace. That framing is wrong, and it is costing you the single most consequential decision you will make about AI. The technology does not decide what it does to your workforce. You do, through how you choose to deploy it.

The scale is real, which is why the framing matters so much. Brookings research finds that more than 30% of all workers could see at least half of their occupation's tasks significantly disrupted by generative AI. And unlike previous waves of automation, which fell on routine and manual work, this one reaches cognitive, non-routine tasks concentrated in middle- and higher-paid professions. The disruption is genuine and broad. But the same body of research makes a point most coverage skips straight past: the effect on workers is not predetermined. It depends on choices, on how the technology is designed and deployed, and on whether the people doing the work have any hand in it. The wave is coming; where it breaks is up to you.

If the disruption is coming anyway, what is actually a choice?

The most important one: augment or automate. Point AI at a role and you can go two ways with the identical tool. You can use it to take the routine, low-value parts off a person so they spend more of their time on the work that actually needs a human, which makes them more capable and more valuable. Or you can use it to strip the role down to its cheapest automatable core and remove the person. Both are available from the same model. They are not technological outcomes; they are management decisions, and they lead to opposite places. Augmentation keeps capability, morale and the judgement you will still need, and it compounds. Automation banks a one-off saving and leaves a thinner organisation behind. The distribution of who gains, the company, the customer, the worker, follows from that call, not from the software.

Why does the augment-or-automate choice matter beyond being humane?

Because it is usually also the better business decision, and because letting it default is the real danger. Left to drift, deployment tends toward automation and cost-out, since that is the saving you can see this quarter. But the cheapest version of a role is rarely the most valuable one, and hollowing out capability to bank a saving tends to cost more than it saves once you need the judgement you removed. How you deploy also decides whether people lean in or resist: augment openly and your workforce helps the technology land; automate by stealth and they quietly withhold the knowledge that would have made it work. This is the same pattern behind why most organisations fail at AI adoption: the tool is rarely the constraint, and the human system you deploy it into is.

How do I make the choice on purpose?

By treating the workforce outcome as a decision to be designed, not a side effect to be discovered.

1. Decide augment-or-automate role by role. Make it an explicit call for each role AI touches, rather than letting the default (cost-out automation) make it for you.
2. Point AI at tasks, not people. Automate the task and redeploy the person to higher-value work. That is the augment path, and it is a choice you have to actively take.
3. Be honest where automation genuinely wins. Some work is pure commodity, and pretending otherwise helps no one. The point is to be selective and deliberate, not blanket.
4. Involve the people doing the work. They know which parts of a job to hand over and which to keep, and involving them builds the trust that makes any deployment actually land.
5. Own the distribution of the gains. Decide on purpose whether freed capacity goes to growth, to upskilling, or to margin, rather than letting it fall wherever the org chart happens to tip.

What does this change for me as a leader?

It moves AI's workforce impact out of the weather report and into your decision rights. The temptation is to treat it as something happening to you, to be predicted and survived. The reality Brookings documents is that the outcome is unusually open, and open questions are decided by whoever bothers to decide them. If you do not choose deliberately, the default will choose for you, and the default is the cheapest, thinnest version of your organisation.

The leaders who come out ahead will not be the ones who best forecast the wave. They will be the ones who chose, on purpose, to deploy AI so that it made their people more valuable rather than more disposable, and who captured the gains deliberately instead of letting them leak. That is the quieter half of the end of business as usual: the technology sets the possibilities, and the leadership decides which of them actually happens. On your workforce, it is the most consequential AI decision most leaders never realise they are making by default.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.