Learning to live with AI – researchers map out opportunities and pitfalls ahead

Artificial intelligence, algorithmic management and monitoring systems are among the techniques which labour organisations, academics and technology specialists believe need new regulatory approaches to reflect their increasing impact on the world of work.

In the UK, the Trades Union Congress says that the world of employment has changed dramatically since the existing policies on data were published a decade ago.

It now wants to see government guidance “significantly expanded” to cover all areas of the employment relationship where data plays a role.

The TUC expressed its concerns when the Information Commissioner, the UK’s independent body set up to uphold information rights, recently asked for views on employment practices.

According to Matt Creagh, policy officer at the TUC, workforce data is increasingly being used to make “unfair, unsafe and discriminatory decisions” about workers.

“Work has changed significantly since the last code was published,” he said in a blog post. “New types of data-driven work have emerged. For example, algorithms determine the pay rates and workloads of workers in the gig economy.”

He also points out how new technologies have emerged that can harvest vast quantities of workforce data.

“[They] can be used by managers who have staff working remotely,” he explains. One productivity tool, for instance, “logs the hours staff work, counts the number of keyboard strokes made in an hour, records social media usage, and takes photographic ‘timecards’ every ten minutes via a webcam”.


Report recommends worker involvement in algorithm design


The TUC is not the only body in the UK with concerns. The All-Party Parliamentary Group (APPG) on the Future of Work, for example, this month published its report New Frontier: Artificial Intelligence at Work. Among other things, it recommends that workers should receive a full explanation of the purpose, outcomes and impacts of algorithms. This would include access to metrics used to monitor, allocate work, pay workers, and discipline them. The APPH also recommends a new right for all workers to have reasonable involvement in the design and deployment of algorithms that are likely to have significant impacts on them.

This is all familiar territory for Peter Cappelli, the George W Taylor Professor of Management at The Wharton School of the University of Pennsylvania in the United States.

He has analysed the rise of algorithmic management for a number of years and has seen how businesses such as platform services have embraced the benefits this approach can provide.

“If we think about where data science and algorithms have been most prominent, it is in creating electronic platforms for matching independent contractors and clients, especially ride-sharing businesses,” he explains.

He believes there are a couple of reasons why this is the case.

The first is due to the conceptually simple nature of the task: you match one party to another based on simple criteria, such as how close they are to each other.

Another reason relates to the law, where some companies may hope that using technology to manage workers will avoid the appearance of an employment relationship. “To prevent supervising the contractors and turning them into employees, it is safer to do it through electronic means, compensation nudges and algorithms where no humans are involved,” he explains.

The gig economy, he suggests, is a natural fit because they are such simple structures. As a result, the data science aspects are a big part of what they do on a daily basis.

“Most other businesses are far more complex,” he says. “They may use lots of algorithms and other tools, but they are less at the core of what they do.”


Laws will catch up eventually, so companies should plan for that


The various legal challenges to algorithmic management that have been issued in different jurisdictions are an expected part of the process, according to Cappelli, who argues that laws always lag behind new developments.

However, he also warns that laws potentially have the power to fundamentally change these operations – possibly even shut them down. It’s why he believes the onus is on companies.

“They should always be aware of what they are doing and the risks involved,” he says. “There are always risks, but the companies here will have made a decade worth of business before legal changes come.”

Looking to the future, Cappelli believes it will take time for everything to settle down, with contexts established in which these algorithmic solutions work. “I expect slow and modest increased use of these practices over time,” he adds.

Mark van Rijmenam, founder of Netherlands-based technology consultancy Datafloq and author of the book, The Organisation of Tomorrow, predicts artificial intelligence will become increasingly important to businesses.

In fact, he argues that the sheer amount of data being collected about employee performance and interaction with customers means the workforce can be run by algorithmic managers.

“There are already examples of dark factories with no lights because they are completely run by artificial intelligence and robotics,” he says. “AI is going to take over more and more jobs.”

The benefits to businesses in platform services and other sectors, he argues, are clear. The first is lowering the organisation’s expenses by offloading some of the management costs to artificial intelligence.

The second concerns efficiencies. AI can schedule employees, allocate tasks effectively and monitor performance. At its best, this will result in enhanced levels of productivity.

“In theory, it offers lower costs, greater efficiencies and faster decision-making,” says van Rijmenam. “However, the downsides are significant.”


The Big Brother imbalance


He cites three main problems with algorithmic management: surveillance and the data needed to manage staff; a lack of transparency; and the subsequent feelings of dehumanisation.

“There’s a power imbalance between the organisation and gig workers who don’t know the extent to which they’re being monitored, how their data is used, or how the algorithms work,” he adds.

He wants to see a fresh approach in which humans and artificial intelligence are working in harmony. It’s a collaboration he expects to see happening a lot more.

“You would have the speed and efficiency of AI, as well as the human components that are important to make a decision within organisations,” he adds. “I think is crucial because we are social creatures and need to be able to interact with each other.”

This is where he believes the true potential of AI can be found. “It has the potential to make organisations more humane,” he explains. “If it’s being implemented in the right way it can take over the mundane tasks and enable a stronger human connection.”

– Rob Griffin PlatformsIntelligence contributing writer

Photo: Gerd Altmann

Print Friendly, PDF & Email