Obviously the "gig economy" was ground zero for this bullshit, with delivery drivers/riders and rideshare leading the pack, followed by all the other jobs getting sucked into piecework: dog walking, nursing, house cleaning and more. But - as the report notes - 79% of EU companies are doing algorithmic management.

2/

The report (written by freelance writer Ben Wray) is published a year after the EU's Platform Work Directive was passed, and a year before it will go into force in all 27 EU member states. It's doing several different jobs: capturing the extent and abuses of algorithmic management; describing how workers can fight back; connecting this to the new EU law,.

3/

It makes the case for unions to invest heavily in making use of the Platform Work Directive's provisions to transform the EU labor market and protect the vast majority of EU members (not just those in unions) from bossware in all its forms.

Algorithmic management poses serious challenges for trade unions. It gives bosses a massive information advantage over workers at the bargaining table, capturing fine-grained information about activity on the shop floor.

4/

It creates opportunities for bosses to violate collective bargains on a per-worker basis, changing the work conditions and pay for every worker and even for every job. It lets bosses spy on workers even when they're not on the clock, and offers many ways for bosses to retaliate against workers.

5/

Workers trapped by algorithmic management are stripped of agency and problem-solving opportunities. They are put under relentless time-pressure and can be forced into dangerous situations (as when a gig delivery app insists that riders follow a prescribed route, even when accidents and other hazards are in the way).

6/

"Cloudworkers" and other algorithmically managed workers are relentlessly surveilled. Platforms like Upwork can switch on workers' device cameras and photograph them while they work. Often, worker data is sold to data-brokers and other third parties. Biases in gig platforms' algorithms can victimize workers - Black workers, for example, are sometimes fired by apps after failing a facial recognition step (facial recognition works less reliably with darker skin-tones).

7/

The app accuses the worker of violating terms of service by sharing their accounts and kicks them off.

Of course, there's no appeal for this. Algorithmic management goes hand-in-hand with other high-handed measures, like replacing the HR department with a chatbot or a semi-attended info@ email address. You also can't reach the HR department when your pay packet is light, facilitating wage-theft.

8/

When payment systems fail, workers are sometimes left with the bill for their robo-boss's technological failures.

So it's quite important that unions figure out a strategy to address algorithmic management. That's where the Platform Work Directive comes in. The PWD has quite sweeping and bold provisions that can protect workers, but these new rules aren't self-enforcing. Many EU states' data commissioners are grossly underfunded and stretched thin.

9/

While the PWD grants workers many rights, they will need to demand those rights - on the job and in the courts.

The new Directive, in combination with the General Data Protection Regulation (the EU's existing privacy law), allows workers and their representatives to demand extensive data-sets from employers, documenting everything from the algorithmic decision-making that goes into firing workers from an app to the process of calculating their pay and beyond.

10/

But employers deliver this data in obfuscated, hard-to-parse formats. Wray advocates for unions to staff up their own data analysis groups that can assist in these requests and make sense of the results.

Wray also advocates for union technologists who can produce worker-side apps that monitor boss's apps - like the UberCheats app, which compared the mileage that Uber paid drivers for to their actual distances traveled.

11/

While it's important for workers to be able to access the information their bosses have amassed on their work and personal lives, it's just as important that workers not be limited to working with data that bosses are willing to hand over. Employers can't be trusted to mark their own homework.

By investing in technology, unions can close the information gap with employers, and even use data and apps to gain an advantage over bosses.

12/

Wray describes how gig workers created "counter apps" that documented wage-theft, enabled mass refusal of lowball offers, and helped workers win their rights in court.

This technological capacity can also help union organizers, providing a unified digital back-end for union drives in all kinds of shops.

13/

Wray acknowledges that it might be hard for unions to do advanced tech work in-house from the jump, and he isn't adverse to having some of this work contracted out to third parties. But he proposes that this kind of arrangement should be modeled on "Chinese industrial policy...which in the 1980s and 1990s was known for bringing in western technological expertise but ensuring that it was the Chinese state and Chinese companies that reaped the knowledge from external experts."

14/

He also moots the possibility of several unions combining forces to create a joint workers' technology shop that develops and supports tools for all kinds of unions across Europe. This sounds like a very exciting idea indeed - and maybe the answer to the legion of programmers who've asked me repeatedly how they can use their technical skills for good.

15/

And as mentioned, the GDPR offers broad powers for workers to push back against bossware abuses. It lets workers demand the ratings system used to assess their work and to demand corrections to their scores - and it bans "hidden internal evaluations" of workers. It also gives workers the right to demand human intervention in automated decision-making.

16/

When workers are "de-activated" (kicked off the app), the GDPR lets workers file a "subject access request" that forces the company to divulge "all personal information relating to that decision" with workers having the right to demand corrections to "inaccurate or incomplete information."

17/

Despite the breadth of these powers, they have rarely been used, largely thanks to some rather gaping loopholes in the GDPR - for example, bosses can use the excuse that divulging information would reveal their trade secrets and expose their IP. The GDPR limits how far these excuses can go, but bosses routinely ignore those limits. Same goes for the all-purpose excuse that the algorithmic management is delivered by a third party tool.

18/

This excuse is illegal under the GDPR, but bosses roll it out all the time (and get away with it).

The Platform Work Directive patches many of the defects in the GDPR. It bans processing "a worker’s personal data in relation to: their emotional or psychological state; private exchanges; when they are not using the app...

19/

"...On the exercising of fundamental rights including worker organising; things that are personal to the worker including sexual orientation and migration status; biometric data when used to establish that person’s identity."

20/

It expands rights to examine the workings and findings of "automation decision-making systems" and to demand that those findings be exported to a form that can be sent to the worker, and bans transfers to third parties. Workers can demand their data in a form that can be used e.g. to get another job, and their bosses have to pay any expenses associated with this.

21/

The Platform Work Directive requires strict human oversight of automated systems, especially for things like de-activations. The Directive requires EU member-states to hold hearings every two years on this process. Workers have the right to demand human review of any automated decision, and sets a deadline of two weeks for bosses to reply.

22/

If the platform has made a mistake, it has two weeks more to make it up to the worker, either by giving them their jobs back, or paying "adequate compensation" for damages.

The Directive bans platforms from arbitrarily changing how their back-ends work and requires bosses to notify workers and consult with them on "changes to automated monitoring or decision-making systems." It requires bosses to pay experts (chosen by workers) to assess these changes.

23/

All these new rules are exciting, but they'll only come into force if someone fights when they're broken. That's where unions come in. If bosses are caught cheating, the Directive requires them to reimburse unions for any experts they hire to fight the scams.

24/

Wray proposes a detailed series of recommendations to unions for things they should demand in their contracts to maximize their chances to capitalize on the opportunities afforded by the Platform Work Directive, such as establishing a "governance body" within the company "to govern data formation, storage, handling and security issues. This body should include shop stewards and all members of the body should receive data training."

25/

e also sets out technological tactics unions can fund and capitalize on to maximize their use of the directive, such as hacking apps to allow gig workers to increase their earnings. He writes warmly of "the sock-puppet method," where many test accounts are used to place and book work through platforms to monitor their pricing systems to detect collusion and price rigging. This has been successfully used in Spain to create the basis for an ongoing lawsuit over price collusion.

26/

The new world of algorithmic management and the new Platform Work Directive offers many opportunities to organized labor. However, there is always the possibility that an employer will simply refuse to follow the law - as Uber has done, after it was found guilty of violating data disclosure work and was fined €6,000/day until it came into compliance. Uber's now paid €500,000 in fines and has not disclosed the datat that the law and the courts require of it.

27/

With algorithmic management, bosses have figured out new ways to evade the law and steal from workers. The Platform Work Directive gives workers and unions a whole suite of new tools to force bosses to play fair. It's not going to be easy, but the technological capacity workers and unions develop here can be repurposed to wage all-out digital class warfare.

28/

Image:
Cryteria (modified)
commons.wikimedia.org/wiki/Fil…

CC BY 3.0
creativecommons.org/licenses/b…

eof/

File:HAL9000.svg - Wikimedia Commons

^{commons.wikimedia.org}