The Performance Case for Being Human: Framing our Report from the Mayor's Taskforce on AI and Jobs

London now has a serious plan for governing AI at work. What it asks of policy, it also asks of you.

On 21 July the Mayor of London’s AI and Jobs Taskforce publishes its report, and I’d encourage you to read it in full. I served on the taskforce as its only psychologist, and I’ll say at the outset that I respect the work and I respect Martha Lane Fox’s leadership of it. It is a serious attempt to give a city agency over what AI does to work, rather than to react once the damage is visible and hard to undo.

The organising idea is a move from adaptation to agency, and it is the right one. Most of the public conversation treats AI as weather, something that happens to people, with the role of government reduced to handing out umbrellas after everyone is soaked. The report rejects that. It argues that AI adoption is the sum of millions of decisions made by employers, educators, investors and workers, and that those decisions can be shaped rather than merely absorbed. On that foundation it builds three civic capabilities: seeing change early enough to act, shaping how work is redesigned as it arrives, and creating new routes into good work as older ones close. The move I’d single out is the loop that ties the three together, because early intelligence feeds the work of redesign, redesign opens new pathways, and what happens in practice feeds back into the intelligence. That is what makes the whole thing adaptive rather than a shelf of disconnected programmes, and it is the most distinctive thinking in the document.

Here is the thread I want to pull, because it runs straight from the report to your desk. The report is clear that the decisions which matter are made inside organisations, in the room where someone decides which tasks to automate, how a role is redesigned, and whether the people affected are consulted or simply told. Policy can set the incentives around that room, and it cannot enter it. That half of the work belongs to leaders, and the report has handed you the scaffolding to do it well. So let me set out what now falls to you.

Start with the people you employ. The report asks employers to redesign work so that AI augments people rather than replacing them, and that is not a soft preference. It is the point where the performance and the ethical case turn out to be the same. Work gives a person an income, and it offers them a good deal more besides: a shape to the day, a place among colleagues, a status they recognise in themselves, competencies that can be cherished even as they evolve, and the sense of doing something that counts.

Take the task away without redesigning the role, and you have not simply removed a cost line. You have cut someone loose from part of who they are. APS Intelligence’s AI Orientation Continuum, the framework my team designed to support effective AI integration, describes the distance between using these tools as a partner to your judgement and surrendering the judgement to them, and that distance is exactly where identity is either protected or quietly dismantled. Your duty of care does not pause because the disruption arrives as software. It sharpens, because this kind of harm is easy to miss until it has already happened.

The report asks employers to make commitments, and a commitment is only worth the behaviour a leader will hold to when holding to it costs something. Set the floor yourself. Decide out loud what you will not do to your workforce in the name of adoption, and then keep to it when the quarterly numbers make keeping to it inconvenient.

The second scale is your organisation. Sustainable performance and the cheap version of AI adoption are not the same thing, and several firms have already learned the difference the expensive way, cutting people, watching quality and trust fall, and reversing course within the year. Adoption succeeds or fails far more on the quality of management and the conditions around it than on the capability of the model itself. You also carry an obligation that rarely reaches the adoption plan, which is an obligation to the place your organisation calls home. When you restructure, the consequences do not stay inside your walls. They land in the boroughs your people live in, in the entry-level role a young Londoner needed as a way in, in the tax base and the high street. The report is right that London’s real strength is its ability to bring employers together around problems none can solve alone, and your organisation is one of those essential conscious employers whether or not it chooses to behave like one.

The third scale is the one it is tempting to leave to economists or therapists, and I would ask you not to. When large numbers of people are set adrift from a professional identity, the effect does not stay economic. People who no longer recognise themselves in their work, and who conclude the system has no further use for them, do not quietly retrain and wait their turn. Some of them stop believing the arrangement of ‘good worker for good work’ is worth defending at all.

A city, and a country, can absorb a great deal of change while people trust that the change is being done with them, and it absorbs far less once they decide it is being done to them. That is not a forecast. It is the reason the choices at the first two scales matter well beyond your own accounts.

None of this is an argument against the report. It is the report’s own logic carried through to the desk where the decisions actually get made. Read the report today, and then treat the part that falls to you as the part that was always going to matter most.

What will you do to ensure there remains a mandate for good work, good workers and a social contract that protects the future for us all?

Read the report here: Shaping how AI changes work in London | London City Hall