AI in the Workplace: Efficiency and Control

Subscribe to the weekly posts via email here

Built-in AI features are increasingly part of a variety of applications/tools that are used in the modern workplace. For example, Microsoft has integrated their Copilot AI into both Microsoft Teams (frequently used in office spaces as a communication and administrative tool) and every major application in Microsoft 365 (Microsoft, 2025). However, the individual workers who are the users of these applications/tools are rarely consulted about adopting these features, and may not always want them. In many workplaces, workers do not have a choice as to the tools/applications which they are required to use in order to do their job, as this decision is made at the departmental, company, or sometimes industry level. For example, in most academic or research libraries it is beneficial to maintain access to certain shared library systems (such as the shared metadata provided by WorldCat) which run through OCLC (an American membership-driven library organisation providing shared services) (OCLC, 2025). This means that many library workers must work with OCLC applications/tools, and are not themselves given a choice as to whether or not additional AI features are added.

Employers’ willingness to both adopt and accept built-in AI as a feature in virtually every workplace tool/application is partly drawn from a fixation on efficiency and a desire to save time, as these AI features are marketed and presented as helping to make things run smoother, faster, and better organised for both organisations and individuals (OCLC, 2025).

 Western society and culture, which follow capitalist principles, view time as a commodity. This is a perspective which sees time as a finite resource which should be controlled. Time is perceived as a measure of work completed, a resource to be  exchanged, and a potential bargaining tool (Adam 1995, 87). Most importantly, from this perspective, any time not considered to be used efficiently is assumed to be wasted.

In this context, built-in AI features can be seen as extremely beneficial, with AI particularly useful in saving time on repetitive, everyday workplace tasks (Reyes and Mullender, 2025).

However, AI features may not always be as helpful or efficient as they are presented. In fact, in some instances users, and even organisations, can find themselves spending more time correcting errors or cancelling unwanted suggestions put forward by built-in AI features than they are able to save through their assistance (Goldenfein and Yang, 2025). This can be particularly prevalent in those tools where users are not given the option to switch off any AI assistance. The severity of these issues varies from tool/application to tool/application, and so is more prevalent depending on the job/task.

Also, Sociological studies have indicated that workers find greater meaning, and consider themselves to be at their best, when they feel a sense of control and autonomy over the time they spend in their work (Bailey and Madden, 2017). Although promoted, and intended, as helpers and efficiency aids, there is a possibility that AI assisted tools/applications can contribute to a feeling of the loss of this autonomy, and to a feeling of further loss of control over workers’ own work itself (Walther, 2025).

Still, there is a continued, accelerating push for greater AI integration into workplace tools/applications, and the ultimate effect is yet to be seen.


The Challenge

For your workday, keep a record of every AI assisted tool/application you come into contact with as part of your job, and how you interact with them. For each one, ask:

  • What do they do to assist you? 
  • How accurate are they at doing it? 
  • Do you find them helpful? intrusive?
  • How much control do you have over them? (can you turn them off, etc)

At the end of the day, give each an assessment as to whether or not you feel AI enhances the tool/application or not.


Subscribe to the weekly posts via email here

References



Disclosure

No GenAI was used in either writing or preparing this challenge post. However, several of the applications used had AI assistance included as a feature, whether the author requested it or not.

Your Challenger: Alec Mullender