How algorithms are reshaping organisational control – A new study

Algorithms and algorithmic technologies are probably the fastest growing technologies affecting the world of work today. Algorithms are rapidly becoming a daily part of organisational operation and control. You really need to understand how they are changing the world of work.

In this post we look at some new research about how algorithms are changing organisational functions and control (this is an excerpt from a research briefing sent to members last year). 

Algorithms

Algorithms refer to a defined routine, programme, procedure, process or set of rules and calculations that are followed in sequence to achieve some desired output, problem-solving operations, or other action, usually by computer or technology means.

Algorithms are increasingly used in improving decision-making, allocating and coordinating resources and complex tasks in faster, more reliable and efficient ways than human processing of an action. Algorithms are particularly suited to complex repetitive tasks and have been found to produce significantly greater levels of efficiency, innovation and profitability

Indeed, recent studies have shown that significant proportions of organisational control, processing and tasks have at least some element of algorithm involved in the process.

Organisational control as contested terrain

From a research perspective, organisational control has traditionally been seen as ‘contested terrain’ in which various elements and ideologies within the organisation vie for control over:

1. the tasks and direction
2. evaluation and
3. discipline

of and within the organisation

Leaders, employers and organisations are typically engaged in some form of continuous change/innovation in order to remain relevant within a marketplace and, therefore, need to control the organisational processes and production in order to move the organisation (and its people) in a desired direction.

Three primary forms of organisational control are:

Traditional organisational theory suggests that the three primary forms of organisational control are:

1. Simple or direct control
2. Technical control,
3. Bureaucratic control

Simple or direct control

Simple or direct control refers to efforts by leaders and managers to directly appeal to or direct workers to do things that are consistent with the organisation’s aims and goals. This direct approach tends to work best in smaller organisations with few workers, however, as a workforce grows in numbers other forms of control need to be exerted.

Technical control

Technical control refers to situations where control is exerted through workers having to serve production lines and a range of technologies that make people work at a particular pace, way and tempo to match that of the technology. In effect, technical control of individuals’ working life is embedded in the technical and physical facets of the organisation

Bureaucratic control

Bureaucratic control, on the other hand, is based on structures, systems, policies, procedures and rules that guide people’s actions, behaviours and decisions.

A new study on algorithim

A new study by a team of researchers from the MIT Sloan School of Management, Stanford School of Engineering and Stanford School of Humanities and Sciences has looked at the impact of algorithmic technologies on organisational control.

Findings on algorithm

Firstly, the study found that technological (as opposed to technical) innovation is both ubiquitous and pervasive in just about every workplace. At the heart of most technological innovation are algorithmic technologies, which are essentially making decisions for people in the workplace. Particularly with the advent of machine learning and artificial intelligence, algorithms these days are:

1. Comprehensive, both in the fact that they are involved in many aspects of people’s working life and by virtue of the fact that they collect data about workers, such as biometrics, work patterns, decision-making and a whole range of other data about workers’ decisions and performance.
2. Instantaneous, in that algorithms are continually operating, both in gathering data and in creating outputs – including decisions – and that they tend to operate in real time without the lag of most human/team/board decision-making.
3. Interactive, in that people tend to interact and engage with algorithms in real time and they provide continuous feedback for workers.
4. Opaque, in that it is rare that managers, leaders and workers have the technical literacy required to understand how the algorithms are working or when bias or other problems occur.

…at the heart of most technological innovation are algorithmic technologies

The study found that algorithms tend to involve algorithmic:

1. Recommending
2. Restricting
3. Recording
4. Rating
5. Replacing
6. Rewarding

Algorithmic recommending

Algorithmic recommending algorithms tend to provide feedback to workers and prompt them to make decisions preferred by the choice architecture of the algorithm. Whilst recommending algorithms are considered to augment people’s decisions, in reality workers tend to just apply the decision the algorithm is putting forward.

Algorithmic restricting

Algorithmic restricting, on the other hand, frequently both restricts workers’ behaviours and controls them in a manner which complies to the algorithm. Secondly, however, algorithms often restrict the information and data available to workers, both in terms of the actual data and also in terms of understanding how that data is being manipulated.

Algorithmic recording

Algorithmic recording refers to a situation whereby finely detailed recordings and analysis of workers’ actions and behaviours are conducted, usually to provide some form of feedback and real-time adjustment of the workers’ performance.

Algorithmic rating

Algorithmic rating refers to both worker rating and ranking of tasks, decisions, et cetera and also refers to the rating of employees’ work. This usually feeds into the use of predictive analytics, which predict worker potential, performance, training needs and potential achievements.

Algorithmic replacing

Algorithmic replacing refers to 2 situations:

1. The first form of algorithmic replacing is where algorithms are used to hire and fire workers and replace them with better suited employees, or people who are more likely to follow managerial directives. In this situation algorithms considerably speed up the hiring and firing process.
2. The second form of algorithmic replacing is where algorithms are used for decision-making and task management, thereby replacing the need for human decision-making, management and supervision.

Algorithmic rewarding

Algorithmic rewarding is based on two concepts:

1. A situation where individuals’ pay and reward are defined and controlled by predefined algorithms, or
2. Where individual workers or teams are given “psychological” rewards based on algorithms. For example, the gamification of work, in which individuals can gain points or something similar and where work is structured like a video/computer game.

Previous studies – Finding

A number of previous studies have found that this new form of organisational control (algorithmic control) can lead to the following worker outcomes including:

• Higher levels of dependency
• Frustration
• Bias
• Lower levels of well-being
• Reduced employee voice
• Higher levels of productivity
• High levels of turnover and turnover intention
• Lower levels of human interaction and human oversight
• Precarity or an increased precarious position for the organisation if the mechanisms of the algorithms are not understood within the organisation

The bottom line of the study is that algorithmic control is part of the new contested terrain of organisations. Additionally, this form of organisational control is occurring with little ethical, legislative or health and safety/privacy oversight, as it is largely broken down into small areas of working and, as a result, has been largely hidden from external oversight or consideration.

Reference

Kellogg, K. C., Valentine, M. A., & Christin, A. (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals, 14(1), 366-410.

The Hottest Research Trends 2018: Analysis by The Oxford Review