Summary: Bureaucracy might sound old-fashioned, but actually offers a useful lens for understanding AI’s mixed effects on HR. Used badly, AI creates new bureaucracy but used well it removes friction while keeping human oversight.
The word ‘bureaucracy’ is not always used in workplace headlines and can feel old-fashioned. Yet bureaucracy still exerts a pervasive influence in our governments, organisations and everyday lives. This makes it a powerful, under-utilised perspective for interpreting opportunities and threats concerning AI and HR.
While bureaucracy generates rules, roles, structures and experiences of hierarchy and procedure, ‘generative’ AI produces text, images and code. With agents generating autonomous behaviours, decisions and multi-step actions.
This creates tensions at the AI-bureaucracy interface, as workplace systems are simultaneously refined and expanded.
Bureaucracy is not inherently bad. But as a tool to bring order, fairness and predictability to workplaces, it can radically fail, creating absurdity, inertia and dehumanisation.
On AI, we might ask: How can we ensure AI reflects positive forms of bureaucracy, while minimising/preventing negative forms?
Understanding AI with bureaucracy
One global 2025 study found 77 per cent of employees reporting AI-related productivity gains in the past 12 months. But 37 per cent of time saved is consumed by rework and HR makes up 38 per cent of those with the most AI-related rework. This rework could be termed ‘AI bureaucracy’.
Yet more positive examples, such as IBM’s 2025 deployment of its fully agentic AI system “AskHR”, present AI as a strong solution for explicitly and successfully handling and removing major layers of past bureaucracy – the delays, multiple handoffs, silos and repetitive, high-volume tasks.
Bureaucracy can help make sense of this mixed picture. It long predates generative AI, yet has always existed in a deeply linked relationship with technology.
Bureaucracy provides clear pillars for driving and governing AI, but intent and design can generate stark pros and cons.
How can we ensure AI reflects positive forms of bureaucracy, while minimising/preventing negative forms?
AI and workplace bureaucracy at its worst
Bureaucracy helps describe and integrate ways in which AI leads to very negative outcomes at work. For instance, the term ‘AI workslop’ is now quite familiar, but reflects how AI can feed back into and expand workplace bureaucracy with insubstantial content and time wasted.
Other, related AI-bureaucratic phenomena include excessive ‘bot-sitting’, where employees can spend nearly a day a week unrewarded just making AI work usable, and ‘bot-shitting’, where up to 69 per cent of users admit to passing on AI work they haven’t properly checked or understood.
At worst, AI and workplace bureaucracy not only destroy productivity and value, but intensify work, speeding it up and extending workloads in unwelcome and unsustainable ways.
Unchecked, wider cultures become obsessed with efficiency and productivity in shallow, self-defeating ways enforced by central power and control.
AI and workplace bureaucracy at its best
Bureaucracy at its best, instead of forcing strict compliance, is enabling – helping stakeholders master their tasks and solve problems, reinforcing learning, fairness, transparency and flexibility.
This provides guidance and meaning on the ground, not just the illusion of order from the top or centre.
Examples of this:
- Siemens used agentic AI to streamline workflows across the employee journey
- Chinese finance group Ping An used advanced gen AI models to optimise approvals, removing friction and manual checkpoints
- Unilever used an AI stack to enhance objectivity, efficiency and fairness within high-volume recruitment funnels.
Such examples reflect AI-enabled systems, with carefully designed combinations of automated efficiency and human intervention.
AI bureaucracy at its best helps orchestrate work across silos, interfaces with accountable employees, is overseen by leaders allowing flexibility and empowerment, and capitalises on ‘middle office’ opportunities.
At worst, AI and workplace bureaucracy not only destroy productivity and value, but intensify work, speeding it up and extending workloads in unwelcome and unsustainable ways
The role of HR in addressing AI and workplace bureaucracy
HR can get a bad reputation for enforcing forms of bureaucracy that interfere with ‘real’ daily work. However, in some ways, this is what makes HR well-placed to take some ownership for distinguishing between necessary governance and redundant friction when adopting and implementing AI.
Many HR functions and professionals will be reflecting on how AI relates to transformation and redesign of their roles and operating models.
Business partners, L&D specialists, total rewards leaders and other practitioners can redesign time allocations and workflows to offer more human-centred experiences and services.
HR can help audit processes to prevent digital or algorithmic bureaucracy by limiting unnecessary control and complexity.
HR can help manage processes of AI adoption in training and change management experiences.
Finally, HR can encourage responsible AI use through simple, open policies and two-way communications.
The future of AI and workplace bureaucracy
In general, the future of AI involves critical uncertainties and various differing scenarios.
By contrast, research suggests that bureaucracy will remain a predominant, persistent, defining feature of workplaces and organisations – a ‘necessity we love to hate’. It can support or impede missions, goals and values.
However, as bureaucracy and AI continue to interact with each other, we can imagine different environments we might wish to try to create and work in.
A positive future vision might consist of smooth, fluid, reliable AI automation coupled with human-centric oversight and augmented interactions.
Negative futures invite extreme caution, crises to be avoided. Here, AI fuels ‘hyper-bureaucracy’ – overwhelming, relentless, unnecessary flooding and bloating of systems, burdening workplaces unsustainably, generating huge losses.
A crisis of human trust and agency also looms if we cede control to opaque, rigid, scaled-up systems.
Actionable insights
HR professionals, leaders and stakeholder teams can interrogate strategies and values by linking AI-related changes and impacts to features of bureaucratic organisation.
- Embed rituals and incentives with leaders and teams across levels and functions for removing negative ‘friction’, ‘bottlenecks’ and ‘workslop’ surrounding AI use
- Encourage usage of checklist tools and frameworks that redefine, question and measure whether AI is adding value to particular processes
- Adopt agile work designs, such as ‘guardrails’, that balance top-down ‘minimum viable’ AI guidance with bottom-up empowerment, supporting sensible team purpose, data use and resource allocation
- Engage stakeholder networks and partners for navigating how AI-related change can be understood through bureaucracy in specific industry/sector contexts. Eg. Regulations, digital transformations, hierarchies, functions and risk management.
If you enjoyed this article, why not read another by Professor Tom Calvard?: From automation to augmentation: Upskill your workforce for the AI era
