None of the proposed rules would let workers say no to being monitored. One would only raise the payout in a case you had already won

A rule only matters if someone can enforce it. So the useful question about the monitoring consultation is not what each option says, but how it would reach a worker who is being watched.

Option one: a code of practice

A code of practice would be issued under a 1992 employment law that already allows them. Breaking it would not, on its own, give anyone a claim. It would matter only inside a claim that already exists, such as unfair dismissal or discrimination. If the worker won that claim, and the tribunal found the employer had unreasonably ignored the code, the award could be adjusted by up to 25 per cent.

That route depends on employment status. Unfair dismissal is a claim for employees. People with worker status but not employee status can bring a narrower range of claims, such as discrimination, and the self-employed fewer still. People who work through apps, the consultation says, are typically not employees. It says so itself: some of the people affected would be less likely to benefit.

Option two: a duty to consult

The second option would require employers to consult, and possibly negotiate, before introducing monitoring. The document models it on collective redundancy consultation, where the duty is to talk with a view to agreement, not to reach one. Remedies could be awards for failing to follow the process, not an order stopping the system. The document asks whether a duty should cover employees only, add people with worker status, or reach everyone who is not self-employed.

Option three: guidance

The third option adds nothing enforceable. It would explain the law as it stands.

The protection you already have

For all three, the document points back to data protection law, which already applies to everyone whose information is processed, app workers included. Employers must have a lawful basis for monitoring, collect no more than they need, and carry out an impact assessment where monitoring is intrusive. Where a decision with a significant effect is made with no meaningful human involvement, the person has a right to be told, to make representations, and to ask for a human to look again.

What the EU banned, and the UK never mentions

Since 2 February 2025, the EU has prohibited AI systems that infer a worker's emotions from biometric data such as facial expressions or voice, with exceptions for medical and safety reasons. The UK consultation's list of monitoring types includes biometric access and heart-rate and fatigue monitoring. The word emotion does not appear anywhere in it. None of the three options would prohibit any specific practice.

What it adds up to

Every option in the document is about process: telling workers, consulting them, weighing the code in a dispute. That can do real good. Monitoring introduced in the open is usually monitoring done better. But no option gives a worker, or the workers' representatives, a way to say no, and the people the document says are watched most closely are the ones its processes reach least.