Data-driven group actions: how to get the edge over the gig economy

Group actions are here to stay, but if law firms are going to take on the gig economy, they need the tools to match. Disputed.io founder and chief executive Steve Shinn explains more.


There is now no doubt about it: group action lawsuits are now well established in the UK. Be it MasterCard v Merricks or Uber BV v Aslam, the last year has seen a spate of ground-breaking collective litigations that are shaping the way we litigate collectively in England and Wales.

And, as the Supreme Court ruling in the Uber group action in February and the Court of Appeal ruling in the Deliveroo group earlier this summer show, people who work in the ‘gig economy’ are increasingly seeking to enforce their employment rights. So much so that Deliveroo is reported to have set aside more than GBP 100 million to cover future employment lawsuits.

My career has been in tech, so I have been watching the rise of the big gig economy employers with both fascination and concern because their business tools are the same as mine: technology, data and machine learning.

At this point, it is important to differentiate between gig economy workers and freelancers. In the raw gig economy people’s work is clearly being directed by tech organisations that set pay and conditions, but do not guarantee work and do their best to ensure their workers never become employees. A freelancer is somebody who works for a rate they themselves set.

While working for a gig economy supplier to generate a second income is a perfect reasonable thing to do, increasing numbers of people are spending all their working hours waiting on gig work and often not getting it.

I do not believe that an unfettered gig economy is a good business model for workers or capitalism in general. Yet the gig economy behemoths are hard to challenge. The whole point of their business model is to use data and technology to offload risk in a way that satisfies regulators.

That meant that, on the face of it, bringing a group action case against a gig economy company can be daunting. Gig economy workers, buy their nature, are a disparate group.

TOO BIG TO SUE (COLLECTIVELY)?

A law firm might well have a lead claimant to represent the group. But at some point (sooner rather than later) it is going to have to contact potentially thousands of possible claimants

You cannot get their names from a trade union, or an HR department. The obvious thing to do is advertise. To reach out and publicise the case. Then, check the identities of each potential claimant, and make sure they are who they say they are.

The obvious way to do this is by using a call centre. That means that a real person needs to be available to manually process each claimant. Call centres cost money. And they are a cost that comes out of the overall claims pot.

This is sustainable for a small-to-medium-sized claim for a relatively large amount of money, it is a sustainable solution. But if there are a very large number of potential claimants, the costs will spiral.

Unfortunately, that means that the more widespread the abuse, the more expensive the case becomes to litigate. Using the call centre method is likely to mean that some large cases become too expensive to litigate.

Some companies, on the face of this, seem too big to (collectively) sue.

Unless lawyers start to embrace the same tools used by the gig economy employers: machine learning and artificial intelligence (AI).

MACHINE LEARNING AND AI FOR LAWYERS

Many new economy/gig economy employers use AI and machine learning platforms to on-board, check up on, and ultimately pay their employees. It allows them to reduce their recruitment costs and significantly increase recruitment speed, allowing them to process large amounts of people quickly. It is an important factor in the viability of their business models.

The employer creates a platform tailored to their own requirements. Potential employees upload information such as identity data and work permission status. The platform cross checks and verifies it and asks the various necessary questions using AI to guide the information requested. Ultimately, the platform will take bank details for wage payment.

This is now a standard process for many companies, and lawyers should consider using it, in a slightly adapted form, for class building.

They too can set up an AI driven platform that requests and checks all the necessary data.

So, for example, in the case of a taxi company claim, the platform can check that a driving licence ID photograph matches the name of the claimant, that the private hire permit is valid, and that the driver has driven for the requisite number of years to qualify for the claim. It can be tailored to perform all sorts of complex verification processes necessary for the case.

If data is found to be lacking, the system can red-flag the file to a human case handler, who can take it from there. In a data-driven world, the majority of cases should be processed via the platform without human intervention.

The only files that will have to be checked by hand are those that do not meet the criteria.

Efficient onboarding is only the beginning. From then onwards an AI platform can automate the entire process of the claim. It can contact claimants with requests, check the responses are correct and valid, and, ultimately, pay them out.

MERRICKS V MASTERCARD: THE PROBLEMS OF SUCCESS

AI-driven platforms are not just useful for gig-economy cases. They could have a place in many large actions.

Take, for example, Merricks v Mastercard. The Competition Appeal Tribunal certified it last month, which means that, barring a settlement, it is likely to go to trial.

The lead claimant, Walter Merricks, is suing Mastercard for around GBP 10 billon arguing the firm’s interchange fees broke European Union law. His lawyers argue that any individual who bought goods and services from a United Kingdom business that accepted Mastercard between 1992 and 2008 is entitled to a pay-out. That might be as many as 46 million people.

Basic maths shows that, if GBP10 billion is awarded, the average pay out will be £217. Except it will not, because the case costs are paid for from the award before it is distributed.

The lawyers will likely claim some of that in fees, but not enough to make a big difference. However, distribution costs using a call centre are likely to be high.  A distribution platform could reduce the this cost significantly.

This is important because anybody who has been following the judgements in Merricks as it has gone from the CAT to the Supreme Court will be aware that the judges are keen to check that any group action that gets certified is viable. 

They want to ensure that the purpose of the case is to compensate claimants, not just keep lawyers busy. A case that is so expensive to run that it leaves claimants with very little left will not get certified. 

Creating efficiencies in distribution will increase payments to claimants and stop some cases from becoming too costly to be worth bringing.

FUTURE IMPERATIVE

Machine learning and AI have created a quiet business revolution in recent years. It has enabled businesses and individuals to successfully and economically manage large amounts of data.

This is because technology now allows computers to ask much more than simplistic yes-or-no questions.

What AI will not be able to do any time soon is to replace a team of skilled lawyers building a complicated case on behalf of their clients. But increasingly AI is likely to create many tools that lawyers can use to streamline their work processes and, ultimately, increase access to justice.


Previous
Previous

Litigation funding is now for everyone

Next
Next

How Fieldfisher is using technology to drive down costs in volume claims