By Thomas KlikauerJ
uly 11, 2024
Source: Originally published by Z. Feel free to share widely.
![](https://znetwork.org/wp-content/uploads/2024/07/14391226325_f4829d4315_k-600x450.jpg)
janneke staaks - Research Data Management. Flickr.
Ascertaining the role of trade unions is crucial for the debate on algorithmic management. This cannot be restricted to online platforms as used in delivery services, for example.
Instead, when algorithmic management is introduced into traditional companies, it links corporate managers, workers, and trade unions in a fundamentally new way. Perhaps in all of this, three elements are crucial:Big Data: firstly, there is data production which is known as “big data”. Behind the “big” name of big data simply means a lot of data. That is pretty much it. This is a huge volume of data that can also be a complex data set. In general, these are data from a new data source. Almost by definition, big data are data that are more voluminous than traditional data. And oftentimes, conventional processing softwares fail to manage such huge sets of data.
Data Interpretation: secondly, there is data analysis which usually involves human interpretation. In short, it is a human person that makes sense of these newly produced data. For trade unionists, it means that they should not fall into management’s trap of selling those new data as independent, scientific, and objective – despite their appearance of being impartial and unbiased. They are management data, produced, analyzed, interpreted and applied by management for a purpose. In short, they “can” be Weapons of Math Destruction. Worse, they can be used against workers and trade unions.
The Context: thirdly, and as always in labor relations – big data are not produced, analyzed, and used in an organizational vacuum. Instead, they are used in a business organization in which organizational politics exist. This is the unavoidable context in which algorithmic management takes place – in “any” workplaces.
To understand all this, one needs to move significantly beyond the much-trumpeted idea that algorithmic management is mostly about the “surveillance” of workers. As important as surveillance capitalism might be, neither capitalism nor workplace relations are based on surveillance alone.
For one, labor relations at both business organizations and non-business organizations, such as not-yet privatized hospitals and schools always take place inside a regulatory framework that restricts, regulates or, at least, shapes what is known as the digital monitoring of workers.
Instead of overplaying the issue of surveillance, managerial goals are not always – not even under the despotism of algorithmic management – unilaterally oriented towards creating the next super-Orwellian managerialist surveillance regime.
Instead, the outcome of algorithmic management, nevertheless, is likely to be a further centralization of knowledge and power in the hands of authoritarian management. Simultaneously, algorithmic management can, and indeed often does, lead to the disempowerment of trade unions and workers.
Since a few years, algorithmic management has become a crucial topic on debates about the digital transformation of work, algorithmic management, and the role of trade unions and workers in all of this.
Yet, algorithmic management is increasingly used for organizational tasks such as performance evaluation and the disciplining of workers.
Necessarily, one needs to extend the concept of algorithmic management beyond that of an online platform economy and beyond that of monitoring digitally in conventional enterprises.
Overall, algorithmic management is a new technical infrastructure that alters the character of management. Simultaneously, the process of negotiating algorithmic management between management and trade unions has not been fully understood.
All too many people assume that management is using algorithmic technologies simply for watching, surveilling and, subsequently for disciplining workers with management and workers being the only players. This is not always the case.
Worse, algorithmic management does not have a black box against which trade unions’ resistance can do nothing about. Despite the, at times, overemphasis on surveillance, algorithmic systems can still be used in workplaces to record every movement of workers.
In other cases, algorithmic systems are employed by management for – as management often calls it rather euphemistically – “the optimization of work processes” (read: work intensification).
The key to understand what management is, is not that management carries out the surveillance of workers. The only reason why management exists in the first place is that management contributes to profit maximization.
The surveillance of workers is merely a by-product of creating profits or as the language of Managerialism calls it: creating “shareholder value”. To camouflage the truth about management, one of management’s finest, once claimed that, “shareholder value is the dumbest idea in the world”.
A very common metaphor presents algorithmic management as an Orwellian panopticon. Such a Big-Brother-like “I am watching you” cements corporate information asymmetries and strengthens managerial control over the labor process – a kind of Orwellian fascism.
In many respects, real fascism is not that much different from an algorithmic dictatorship. Both, political fascism in society as well as managerial-algorithmic dictatorships (MAD) in companies rely on the mobilization, blind compliance, and utilization of people (in a society) and of workers (in a company).
The former can lead to authoritarianism or fascism while the latter assists what is today known as algorithmic management.
Set against such a blind compliance, workers and their trade unions can use their knowledge as a power resource in collective bargaining negotiations about the introduction and operation of algorithmic management.
Such management-union consultations and collective bargaining negotiations can focus on management-vs.-union bargaining over work tasks and roles that are set to be altered in the wake of algorithmic management. This might also include strategies directed towards what became known as organizational misbehavior.
Beyond organizational sabotage and misbehavior, such union collective bargaining negotiations should cover the step-by-step implementation of algorithmic management tools in still conventional-organizational settings such as, for example, warehouses, office work, delivery businesses, marketing, manufacturing, and even consultancies.
Almost self-evidently, algorithmic management is reaching more and more conventional companies. Worse, algorithmic management not just “builds on” long established Taylorist practices, but it also has the potential to significantly crank-up and actually worsen Tayloristic work regimes.
There are many longstanding traditions of despotic Taylorist performance management regimes in conventional companies. There are, however, also recognizable differences between different workplaces, different forms of platform work, and the application of different forms of algorithmic management.
With the rise of algorithmic management, a digital data-based infrastructure is established. This creates new competences to operate algorithmic systems.
Hence, there is a growing importance of managing such data by organizations. In other words, “data production” remains crucial. In contrast to a simple App on a phone, most algorithmic systems cannot simply and easily be applied by management on the micro-surveillance and Uber-monitoring of individual work processes.
Until the advent of algorithmic management, productivity management in the area of manual labor and blue-collar work, for example, relied mainly on traditional concepts like Taylorist time and motion studies. Later, it was cranked up by lean production and work process optimization – controlling managers with stopwatches on hand.
That changes under algorithmic management. Yet, efforts to apply algorithmic management can run up against pre-existing forms of organizational hierarchies, workplace bargaining, and external labor regulation by the state.
Given this, digital or algorithmic monitoring of work can collide with a range of normative, legal, organizational-bureaucratic, and other managerial practices.
In any case, many traditional workplaces generally introduce “only some” elements of algorithmic management – often, these are “not” the most widely condemned forms like those restricting worker behavior, the sanctioning of workers, and for dismissal.
Still, algorithmic management system can – and often do – alter organizational power relations between management, IT-engineers, workers, and trade unions. This occurs during the entire phase of introducing algorithmic management which is often done in stages:The initial stage is that of creating “goals” (read: management goals – not the goals of workers). At this stage, collective bargaining negotiations focus on what management likes to call “shared” (read: their) objectives for the implementation of algorithmic systems.
In the second step, the production data takes center stage. This process also involves measuring instruments. These are installed and measuring takes place.
Thirdly, data analysis and the “human” (read: manager) interpretation of newly generated and rather vast amount of data that have been produced occur. This is the ultimate goal of management. It leads to what managers like to call “optimization” (read: work intensification).
In all this, the data and increased options to control workers that are generated by algorithmic management also raises a dilemma for management. This is Foucault’s knowledge-power dynamic. This might challenge the authority of management.
In this dilemma, management depends on “other (non-managerial) actors” like IT-workers to produce and make sense of the data that algorithms have produced. This creates the following dilemma for management: On the one hand, management depends on the involvement of IT-workers in developing knowledge from algorithmic systems.
On the other hand, management can use digital tools to centralize knowledge in their hands in order to bypass workers, which is, ultimately, next to impossible.
In any case, the introduction and eventual operation of algorithmic management can alter the significance of knowledge and that can re-shape existing power relations in a company.
Algorithmic management systems can also fortify existing information asymmetries between management and workers/unions.
Ultimately, algorithmic management always depends on knowledge contributions from workers and that creates spaces for organizational resistance by workers and trade unions.
Yet, management increasingly uses data generated by external actors or external companies. In that way, managerial control over data production is strengthened. To counter or at least shape this, workers’ representations like European style works councils and trade unions can force management to negotiate these setups.
Most importantly, management is never a monolithic and single actor. There can be divisions within management at the horizontal as well as the vertical level:Horizontal Divisions are divisions between different management functions, such as, for example, between marketing and operations management, between accounting and HRM, between sales and organizational development, between strategic management and day-to-day management, and so on.
Vertical Divisions are divisions between different hierarchies inside a company: between top- middle-, and line-management, for example. Yet they can also be between head office and subsidiaries or a division (in a multi-divisional company, for example).
In any case, all companies have various hierarchical levels of management and different power dynamics among different management areas.
Top management does not necessarily agree with middle-management and shop floor management on the introduction of algorithmic systems. We know that algorithmic management does challenge, in particular, the power of middle management.
Inside companies, algorithmic systems will virtually assure that new groups of IT workers like AI experts and data scientists are increasingly getting more relevant.
In turn, this is likely to diminish the role of classical middle management. Simultaneously, top management may lose control over whether top-apparatchiks need to maintain – or at least appear to maintain – to be “the” master of the rising levels of organizational knowledge that is created by algorithmic systems.
Yet, algorithmic management might also impact on operations management as industrial-process engineers may become ambivalent and insecure about their role inside a company. They might even face what is known as: technostress.
Up until the event of algorithmic management, such engineers have been asphyxiated inside a mode of thinking set in motion by Taylor’s rather un-scientific management.
With algorithmic management, an entire new relationship between Tayloristic engineers and data scientists may well be one that creates tensions and frictions inside companies.
Despite all the integrated technicalities of algorithmic systems, algorithmic management is still a socio-technical, human-created, and political-economical issue. This is filled with power struggles and potentials for resistance.
Despite this, new technologies like algorithmic systems are almost always introduced into profit-making companies.
And these, in turn, operate inside capitalism. In short, the goal of profit-making will guide the introduction of algorithmic systems. It has done so ever since the pin factory (Adam Smith) and the Spinning Jenny (Karl Marx).
On the one hand, algorithmic systems are causing shifts in knowledge and power relations as they reconstruct new organizational knowledge.
On the other hand, the introduction of algorithmic management makes collaboration with an external service provider and/or internal IT-workers necessary to implement algorithmic management.
During the introduction of algorithmic management, organizational relationship often remains ambivalent regarding the organizational power relations between central and middle management.
In any case, it is rather likely that line management might be undermined because of the implementation of algorithmic management.
All in all, top management should support line management in implementing algorithmic management to safeguard the approval of algorithmic management by line management.
Beyond that, the position of workers and trade unions as well as the power relations between workers, unions, engineers and management is often contested during the implementation phase of algorithmic management.
On the side of workers, trade unions need to make sure that workers are aware that the introduction of algorithmic management is always embedded in an existing regulatory framework.
This offers trade unions a series of opportunities to prevent corporate bosses from using algorithmic management for the Uber-monitoring of individual workers and to increase work pressure on workers to perform.
Collective bargaining negotiations between management and trade unions can lead to a strategy that forces management to abstain from downsizing employment and algorithmically enforce managerially set performance targets.
The setting up of institutional arrangements – union-management committees – are an almost necessary precondition for the advancement of a pro-worker arrangements and the enforcement of workplace rights during the introduction of algorithmic management.
In addition, workers also have their own resources particularly during the “data production” phase to gain power. On the downside, however, is the fact that once “data” are collected, all too often management no longer relies on the direct support of workers.
In other words, once management gets what it wants, it can use new data, new knowledge, and new algorithmic systems against workers. This makes corporate “optimization” or work intensification feasible for management “without” any further involvement of IT-workers.
In short, management is forced to reach compromises with worker and unions during the “data production” phase. But – and this is a very serious “BUT” – after that, there often is a strengthening of managerial power, knowledge, and information asymmetries.
This is further enhanced by algorithmic management that reduces the need for worker involvement.
In the end, algorithmic management can very easily give management even more power to act at will and management – throughout the history of labor relations – will use the knowledge and power it gains from algorithmic management against workers. As long as capitalism exists, workers will need to fight this.
ZNetwork is funded solely through the generosity of its readers. DONATE
![](https://znetwork.org/wp-content/uploads/avatars/klikauer.png)
Thomas Klikauer has over 800 publications (including 12 books) and writes regularly for BraveNewEurope (Western Europe), the Barricades (Eastern Europe), Buzzflash (USA), Counterpunch (USA), Countercurrents (India), Tikkun (USA), and ZNet (USA). One of his books is on Managerialism (2013).
![](https://znetwork.org/wp-content/uploads/2024/07/14391226325_f4829d4315_k-600x450.jpg)
janneke staaks - Research Data Management. Flickr.
Ascertaining the role of trade unions is crucial for the debate on algorithmic management. This cannot be restricted to online platforms as used in delivery services, for example.
Instead, when algorithmic management is introduced into traditional companies, it links corporate managers, workers, and trade unions in a fundamentally new way. Perhaps in all of this, three elements are crucial:Big Data: firstly, there is data production which is known as “big data”. Behind the “big” name of big data simply means a lot of data. That is pretty much it. This is a huge volume of data that can also be a complex data set. In general, these are data from a new data source. Almost by definition, big data are data that are more voluminous than traditional data. And oftentimes, conventional processing softwares fail to manage such huge sets of data.
Data Interpretation: secondly, there is data analysis which usually involves human interpretation. In short, it is a human person that makes sense of these newly produced data. For trade unionists, it means that they should not fall into management’s trap of selling those new data as independent, scientific, and objective – despite their appearance of being impartial and unbiased. They are management data, produced, analyzed, interpreted and applied by management for a purpose. In short, they “can” be Weapons of Math Destruction. Worse, they can be used against workers and trade unions.
The Context: thirdly, and as always in labor relations – big data are not produced, analyzed, and used in an organizational vacuum. Instead, they are used in a business organization in which organizational politics exist. This is the unavoidable context in which algorithmic management takes place – in “any” workplaces.
To understand all this, one needs to move significantly beyond the much-trumpeted idea that algorithmic management is mostly about the “surveillance” of workers. As important as surveillance capitalism might be, neither capitalism nor workplace relations are based on surveillance alone.
For one, labor relations at both business organizations and non-business organizations, such as not-yet privatized hospitals and schools always take place inside a regulatory framework that restricts, regulates or, at least, shapes what is known as the digital monitoring of workers.
Instead of overplaying the issue of surveillance, managerial goals are not always – not even under the despotism of algorithmic management – unilaterally oriented towards creating the next super-Orwellian managerialist surveillance regime.
Instead, the outcome of algorithmic management, nevertheless, is likely to be a further centralization of knowledge and power in the hands of authoritarian management. Simultaneously, algorithmic management can, and indeed often does, lead to the disempowerment of trade unions and workers.
Since a few years, algorithmic management has become a crucial topic on debates about the digital transformation of work, algorithmic management, and the role of trade unions and workers in all of this.
Yet, algorithmic management is increasingly used for organizational tasks such as performance evaluation and the disciplining of workers.
Necessarily, one needs to extend the concept of algorithmic management beyond that of an online platform economy and beyond that of monitoring digitally in conventional enterprises.
Overall, algorithmic management is a new technical infrastructure that alters the character of management. Simultaneously, the process of negotiating algorithmic management between management and trade unions has not been fully understood.
All too many people assume that management is using algorithmic technologies simply for watching, surveilling and, subsequently for disciplining workers with management and workers being the only players. This is not always the case.
Worse, algorithmic management does not have a black box against which trade unions’ resistance can do nothing about. Despite the, at times, overemphasis on surveillance, algorithmic systems can still be used in workplaces to record every movement of workers.
In other cases, algorithmic systems are employed by management for – as management often calls it rather euphemistically – “the optimization of work processes” (read: work intensification).
The key to understand what management is, is not that management carries out the surveillance of workers. The only reason why management exists in the first place is that management contributes to profit maximization.
The surveillance of workers is merely a by-product of creating profits or as the language of Managerialism calls it: creating “shareholder value”. To camouflage the truth about management, one of management’s finest, once claimed that, “shareholder value is the dumbest idea in the world”.
A very common metaphor presents algorithmic management as an Orwellian panopticon. Such a Big-Brother-like “I am watching you” cements corporate information asymmetries and strengthens managerial control over the labor process – a kind of Orwellian fascism.
In many respects, real fascism is not that much different from an algorithmic dictatorship. Both, political fascism in society as well as managerial-algorithmic dictatorships (MAD) in companies rely on the mobilization, blind compliance, and utilization of people (in a society) and of workers (in a company).
The former can lead to authoritarianism or fascism while the latter assists what is today known as algorithmic management.
Set against such a blind compliance, workers and their trade unions can use their knowledge as a power resource in collective bargaining negotiations about the introduction and operation of algorithmic management.
Such management-union consultations and collective bargaining negotiations can focus on management-vs.-union bargaining over work tasks and roles that are set to be altered in the wake of algorithmic management. This might also include strategies directed towards what became known as organizational misbehavior.
Beyond organizational sabotage and misbehavior, such union collective bargaining negotiations should cover the step-by-step implementation of algorithmic management tools in still conventional-organizational settings such as, for example, warehouses, office work, delivery businesses, marketing, manufacturing, and even consultancies.
Almost self-evidently, algorithmic management is reaching more and more conventional companies. Worse, algorithmic management not just “builds on” long established Taylorist practices, but it also has the potential to significantly crank-up and actually worsen Tayloristic work regimes.
There are many longstanding traditions of despotic Taylorist performance management regimes in conventional companies. There are, however, also recognizable differences between different workplaces, different forms of platform work, and the application of different forms of algorithmic management.
With the rise of algorithmic management, a digital data-based infrastructure is established. This creates new competences to operate algorithmic systems.
Hence, there is a growing importance of managing such data by organizations. In other words, “data production” remains crucial. In contrast to a simple App on a phone, most algorithmic systems cannot simply and easily be applied by management on the micro-surveillance and Uber-monitoring of individual work processes.
Until the advent of algorithmic management, productivity management in the area of manual labor and blue-collar work, for example, relied mainly on traditional concepts like Taylorist time and motion studies. Later, it was cranked up by lean production and work process optimization – controlling managers with stopwatches on hand.
That changes under algorithmic management. Yet, efforts to apply algorithmic management can run up against pre-existing forms of organizational hierarchies, workplace bargaining, and external labor regulation by the state.
Given this, digital or algorithmic monitoring of work can collide with a range of normative, legal, organizational-bureaucratic, and other managerial practices.
In any case, many traditional workplaces generally introduce “only some” elements of algorithmic management – often, these are “not” the most widely condemned forms like those restricting worker behavior, the sanctioning of workers, and for dismissal.
Still, algorithmic management system can – and often do – alter organizational power relations between management, IT-engineers, workers, and trade unions. This occurs during the entire phase of introducing algorithmic management which is often done in stages:The initial stage is that of creating “goals” (read: management goals – not the goals of workers). At this stage, collective bargaining negotiations focus on what management likes to call “shared” (read: their) objectives for the implementation of algorithmic systems.
In the second step, the production data takes center stage. This process also involves measuring instruments. These are installed and measuring takes place.
Thirdly, data analysis and the “human” (read: manager) interpretation of newly generated and rather vast amount of data that have been produced occur. This is the ultimate goal of management. It leads to what managers like to call “optimization” (read: work intensification).
In all this, the data and increased options to control workers that are generated by algorithmic management also raises a dilemma for management. This is Foucault’s knowledge-power dynamic. This might challenge the authority of management.
In this dilemma, management depends on “other (non-managerial) actors” like IT-workers to produce and make sense of the data that algorithms have produced. This creates the following dilemma for management: On the one hand, management depends on the involvement of IT-workers in developing knowledge from algorithmic systems.
On the other hand, management can use digital tools to centralize knowledge in their hands in order to bypass workers, which is, ultimately, next to impossible.
In any case, the introduction and eventual operation of algorithmic management can alter the significance of knowledge and that can re-shape existing power relations in a company.
Algorithmic management systems can also fortify existing information asymmetries between management and workers/unions.
Ultimately, algorithmic management always depends on knowledge contributions from workers and that creates spaces for organizational resistance by workers and trade unions.
Yet, management increasingly uses data generated by external actors or external companies. In that way, managerial control over data production is strengthened. To counter or at least shape this, workers’ representations like European style works councils and trade unions can force management to negotiate these setups.
Most importantly, management is never a monolithic and single actor. There can be divisions within management at the horizontal as well as the vertical level:Horizontal Divisions are divisions between different management functions, such as, for example, between marketing and operations management, between accounting and HRM, between sales and organizational development, between strategic management and day-to-day management, and so on.
Vertical Divisions are divisions between different hierarchies inside a company: between top- middle-, and line-management, for example. Yet they can also be between head office and subsidiaries or a division (in a multi-divisional company, for example).
In any case, all companies have various hierarchical levels of management and different power dynamics among different management areas.
Top management does not necessarily agree with middle-management and shop floor management on the introduction of algorithmic systems. We know that algorithmic management does challenge, in particular, the power of middle management.
Inside companies, algorithmic systems will virtually assure that new groups of IT workers like AI experts and data scientists are increasingly getting more relevant.
In turn, this is likely to diminish the role of classical middle management. Simultaneously, top management may lose control over whether top-apparatchiks need to maintain – or at least appear to maintain – to be “the” master of the rising levels of organizational knowledge that is created by algorithmic systems.
Yet, algorithmic management might also impact on operations management as industrial-process engineers may become ambivalent and insecure about their role inside a company. They might even face what is known as: technostress.
Up until the event of algorithmic management, such engineers have been asphyxiated inside a mode of thinking set in motion by Taylor’s rather un-scientific management.
With algorithmic management, an entire new relationship between Tayloristic engineers and data scientists may well be one that creates tensions and frictions inside companies.
Despite all the integrated technicalities of algorithmic systems, algorithmic management is still a socio-technical, human-created, and political-economical issue. This is filled with power struggles and potentials for resistance.
Despite this, new technologies like algorithmic systems are almost always introduced into profit-making companies.
And these, in turn, operate inside capitalism. In short, the goal of profit-making will guide the introduction of algorithmic systems. It has done so ever since the pin factory (Adam Smith) and the Spinning Jenny (Karl Marx).
On the one hand, algorithmic systems are causing shifts in knowledge and power relations as they reconstruct new organizational knowledge.
On the other hand, the introduction of algorithmic management makes collaboration with an external service provider and/or internal IT-workers necessary to implement algorithmic management.
During the introduction of algorithmic management, organizational relationship often remains ambivalent regarding the organizational power relations between central and middle management.
In any case, it is rather likely that line management might be undermined because of the implementation of algorithmic management.
All in all, top management should support line management in implementing algorithmic management to safeguard the approval of algorithmic management by line management.
Beyond that, the position of workers and trade unions as well as the power relations between workers, unions, engineers and management is often contested during the implementation phase of algorithmic management.
On the side of workers, trade unions need to make sure that workers are aware that the introduction of algorithmic management is always embedded in an existing regulatory framework.
This offers trade unions a series of opportunities to prevent corporate bosses from using algorithmic management for the Uber-monitoring of individual workers and to increase work pressure on workers to perform.
Collective bargaining negotiations between management and trade unions can lead to a strategy that forces management to abstain from downsizing employment and algorithmically enforce managerially set performance targets.
The setting up of institutional arrangements – union-management committees – are an almost necessary precondition for the advancement of a pro-worker arrangements and the enforcement of workplace rights during the introduction of algorithmic management.
In addition, workers also have their own resources particularly during the “data production” phase to gain power. On the downside, however, is the fact that once “data” are collected, all too often management no longer relies on the direct support of workers.
In other words, once management gets what it wants, it can use new data, new knowledge, and new algorithmic systems against workers. This makes corporate “optimization” or work intensification feasible for management “without” any further involvement of IT-workers.
In short, management is forced to reach compromises with worker and unions during the “data production” phase. But – and this is a very serious “BUT” – after that, there often is a strengthening of managerial power, knowledge, and information asymmetries.
This is further enhanced by algorithmic management that reduces the need for worker involvement.
In the end, algorithmic management can very easily give management even more power to act at will and management – throughout the history of labor relations – will use the knowledge and power it gains from algorithmic management against workers. As long as capitalism exists, workers will need to fight this.
ZNetwork is funded solely through the generosity of its readers. DONATE
![](https://znetwork.org/wp-content/uploads/avatars/klikauer.png)
Thomas Klikauer has over 800 publications (including 12 books) and writes regularly for BraveNewEurope (Western Europe), the Barricades (Eastern Europe), Buzzflash (USA), Counterpunch (USA), Countercurrents (India), Tikkun (USA), and ZNet (USA). One of his books is on Managerialism (2013).
No comments:
Post a Comment