How AI Is Changing the Way We Work
Generative AI is transforming work in ways that go beyond simple automation. In many jobs, employees are increasingly collaborating with AI systems by writing prompts, reviewing responses, checking facts, and deciding whether the output is usable.¹˒³ This shift means that human judgment is still central, even when AI is involved.
Rather than replacing people outright, AI is changing the structure of work. Tasks like drafting, summarizing, brainstorming, and finding information may become faster, but employees remain responsible for evaluating the final result, correcting errors, and making decisions based on context and experience.¹˒⁴ In other words, some of the work is disappearing, but new work is appearing in its place.
There is another change that deserves attention. Employees may increasingly be completing tasks with AI that they previously completed with other people.
A recent field experiment involving 791 professionals found that individuals working with AI on product-development tasks could achieve performance comparable to human teams working without AI. AI also helped workers produce solutions that integrated knowledge across professional functions.⁵ This demonstrates the potential of AI as a powerful collaborative tool, but it also raises an important question about what may be lost when AI begins to substitute for some forms of human collaboration.
Working with another person involves more than producing an output. Employees observe how colleagues approach problems, ask questions, disagree, explain their reasoning, receive informal feedback, and learn from one another. These interactions can be particularly important for junior employees who are still developing professional judgment.
The evidence is not yet strong enough to conclude that AI is broadly reducing workplace mentoring or senior-to-junior feedback. However, emerging research suggests that generative AI is already changing workplace relationships and occupational learning. Research involving junior professionals has identified learning challenges when less-experienced employees are expected to work with and provide expertise around emerging AI technologies without yet having the deeper knowledge needed to recognize novel risks.⁶
This creates an important question for the future of work: if AI increasingly performs the tasks through which junior employees traditionally learned, where will that learning happen instead?
How AI Affects Employee Learning and Skill Development
Entry-level work has traditionally served two purposes. It produces work for an organization, but it also gives new employees opportunities to become experienced employees.
A junior employee writes a first draft and receives corrections. A new analyst completes an analysis and learns what they overlooked. A new graduate makes a mistake, receives feedback, and approaches the next problem differently. Over time, these experiences contribute to the knowledge and judgment that allow someone to recognize when something does not look right.
Generative AI complicates this learning process because many of the tasks that AI performs particularly well, including drafting, summarizing, basic analysis, information retrieval, and producing first versions of work, are also tasks that have traditionally provided opportunities for novice employees to practise.
This does not mean that AI will eliminate entry-level work. Current evidence is still developing, and it would be premature to make that claim. What appears more certain is that the content of entry-level work is changing, and organizations will need to consider whether employees are still receiving enough opportunities to practise the foundational tasks required to develop expertise.
This issue becomes particularly important when junior employees are expected to evaluate AI-generated work. Research on generative AI and metacognition suggests that less-experienced users may be particularly vulnerable to inappropriate reliance on AI because they have less existing knowledge against which to evaluate an AI-generated response.⁷
There is an uncomfortable paradox here: AI can help someone complete a task they do not yet know how to complete independently, but that lack of experience may also make them less equipped to recognize when the AI has completed it incorrectly.
Why Cognitive Ergonomics Matters in AI-Enabled Work
One of the many positive outcomes of ergonomics includes making physical work safer and more comfortable, especially through workstation design, posture, and injury prevention. But as work becomes more digital, ergonomics evolves to consider how technology affects attention, memory, decision-making, learning, collaboration, and well-being.⁸ This broader perspective is often described as digital ergonomics.
Cognitive ergonomics is especially relevant in AI-enabled work. It looks at how people think while doing their jobs and how work can be designed to fit human mental abilities and limits.⁹ In the age of AI, this means understanding not just what tasks people perform, but how much mental effort is required to prompt, check, interpret, and appropriately trust AI-generated information.
The Job Demands–Resources model helps explain this balance. It proposes that jobs contain demands that require sustained effort and resources that help employees accomplish their work and maintain well-being.¹⁰ AI can potentially function as either. It can reduce repetitive work and save time, but it can also introduce new demands through continuous checking, learning, decision-making, and changing productivity expectations.¹¹˒¹²
Does AI Reduce Workload or Simply Change it?
One of the strongest arguments for generative AI is efficiency. An employee may produce a first draft in minutes instead of an hour, summarize a lengthy document almost instantly, or automate an administrative task that previously consumed a significant portion of the workday.
But there is an important difference between making a task easier and reducing workload.
A 2026 systematic review of 30 empirical studies examining generative AI and academic workload found that AI did not simply increase or decrease workload. Instead, researchers described a process of workload transformation. AI improved efficiency for routine and procedural activities while introducing additional responsibilities, including reviewing AI outputs, making ethical judgments, and monitoring quality.¹³
The impact also depended on factors such as the employee’s experience, digital literacy, ability to evaluate AI, organizational policies, training, and technological infrastructure.¹³
There is another possibility that deserves attention. If a task that previously required two hours can now be completed in one, the saved hour does not necessarily become recovery time. Organizations may instead expect employees to complete additional tasks.
At this point, there is not enough evidence to claim that generative AI is universally causing work intensification across workplaces. However, the possibility highlights why productivity and workload should not be treated as the same outcome. An employee can become substantially more productive while simultaneously experiencing the same, or potentially greater, overall work demands.
From an ergonomics perspective, the important question therefore becomes not simply “How much time did AI save?” but “What happened to the time that was saved?”
How AI is Changing Cognitive Workload
One of the biggest changes brought by AI is the shift in mental workload. AI can make some tasks easier, such as drafting a first version of a document or summarizing long text. But it also creates new demands because employees still need to verify accuracy, spot mistakes, interpret the information, and decide whether the output makes sense in context.¹˒²
Research involving 319 knowledge workers and 936 real-world examples of generative AI use found that AI changed where employees directed their critical-thinking effort. Workers shifted from gathering information toward verifying information, from solving problems themselves toward integrating AI-generated responses, and from directly completing tasks toward overseeing them.⁴
This distinction matters because doing less of the original task does not necessarily mean thinking less overall; the thinking may simply occur somewhere else in the workflow.
However, the same study identified a potentially concerning pattern. Greater confidence in generative AI was associated with lower reported critical-thinking effort, while greater confidence in one’s own ability was associated with more critical-thinking effort.⁴
This connects directly to automation bias, which is the tendency to rely too heavily on an automated system even when it is incorrect. Because AI-generated content can sound confident, polished, and convincing, employees may accept information without sufficient scrutiny, particularly when they are busy or working under pressure.¹⁴
AI, Critical Thinking and the Risk of Over-Reliance
But How Do You Spot a Mistake if You Have Never Learned What Right Looks Like?
Trust calibration becomes especially complicated for inexperienced workers.
Experienced employees often have years of accumulated knowledge that help them notice when an answer seems unusual. They may recognize an incorrect assumption, missing context, unrealistic recommendation, or subtle error because they have encountered similar situations before.
A novice may not have that same internal reference point.
Research on generative AI and critical thinking suggests that a user’s existing task confidence and expertise influence how they engage with AI. Lower confidence may encourage greater cognitive offloading, while workers with stronger task knowledge may be better positioned to supervise and challenge AI-generated outputs.⁴˒⁷
This creates another paradox in AI-enabled work. The employees who can potentially benefit most from AI support may also be the employees who have the least experience required to determine when that support is wrong.
The concern is not that employees should never use AI to perform unfamiliar work. The more important question is whether organizations are preserving opportunities for people to develop enough independent knowledge and experience to eventually supervise AI effectively.
Frequently Asked Questions
How is AI changing the workplace?
AI is changing how employees complete tasks, make decisions and collaborate. While generative AI can improve efficiency by supporting writing, research and analysis, it also introduces new responsibilities, including reviewing AI-generated information, verifying accuracy and exercising human judgment.
How does AI affect cognitive workload?
AI can reduce the mental effort required for some tasks while introducing new cognitive demands. Employees may spend less time generating information and more time evaluating, interpreting and verifying AI-generated outputs. As a result, greater productivity does not necessarily mean a lower overall workload.
Can AI affect critical thinking and employee learning?
AI can support learning and problem-solving, but over-reliance may reduce opportunities to practise important skills. This is particularly relevant for less-experienced employees who may lack the knowledge needed to recognize errors in AI-generated work. Current research raises concerns about these risks but does not establish that AI inevitably causes long-term declines in critical-thinking ability.
Why is cognitive ergonomics important when introducing AI into the workplace?
Cognitive ergonomics examines how work can be designed to accommodate human mental abilities and limitations. In AI-enabled workplaces, it helps organizations consider attention, decision-making, mental workload and the balance between AI assistance and human oversight.
In part 2 of our blog we will explore further how AI is affecting today’s workplace.
References
- Bankins S, Ocampo AC, Marrone M, Restubog SLD, Woo SE. A multilevel review of artificial intelligence in organizations: implications for organizational behavior research and practice. J Organ Behav. 2024;45(2):159-182. doi:10.1002/job.2735.
- Soulami M, Benchekroun S, Galiulina A. Exploring how AI adoption in the workplace affects employees: a bibliometric and systematic review. Front Artif Intell. 2024;7:1473872. doi:10.3389/frai.2024.1473872.
- Vaccaro M, Almaatouq A, Malone TW. When combinations of humans and AI are useful: a systematic review and meta-analysis. Nat Hum Behav. 2024;8:2293-2303. doi:10.1038/s41562-024-02024-1.
- Lee HP, Sarkar A, Tankelevitch L, Drosos I, Rintel S, Banks R, et al. The impact of generative AI on critical thinking: self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. In: Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery; 2025. Article 1121. p. 1-22. doi:10.1145/3706598.3713778.
- Dell’Acqua F, Ayoubi C, Lifshitz-Assaf H, Sadun R, Mollick E, Mollick L, et al. The cybernetic teammate: a field experiment on generative AI and teamwork. Organization Science. 2026. doi:10.1287/orsc.2025.20702.
- Pachidi S, Berends H, Faraj S, Huysman M. Novice risk work: how juniors coaching seniors on emerging technologies such as generative AI can lead to learning failures. Inf Organ. 2025;35(1):100559. doi:10.1016/j.infoandorg.2025.100559.
- Tankelevitch L, Kewenig V, Simkute A, Scott AE, Sarkar A, Sellen A, et al. The metacognitive demands and opportunities of generative AI. In: Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery; 2024. doi:10.1145/3613904.3642902.
- Antonaci FG, et al. Workplace well-being in Industry 5.0: a worker-centered systematic review. Sensors (Basel). 2024;24(17):5473. doi:10.3390/s24175473.
- Wong SW, Crowe P. Cognitive ergonomics and robotic surgery. J Robot Surg. 2024. doi:10.1007/s11701-024-01852-7.
- Demerouti E, Nachreiner F, Bakker AB, Schaufeli WB. The job demands-resources model of burnout. J Appl Psychol. 2001;86(3):499-512. doi:10.1037/0021-9010.86.3.499.
- Hessari H, Bai A, Daneshmandi F. Generative AI: boosting adaptability and reducing workplace overload. J Comput Inf Syst. 2024. doi:10.1080/08874417.2024.2417672.
- Çögenli MZ. Generative AI as a conditional job resource under job demands in academic knowledge work: directed content analysis using the job demands-resources framework. Front Artif Intell. 2026;9:1774525. doi:10.3389/frai.2026.1774525.
- Su F, Beh WF, Cheah KSL. Efficiency or intensification? A systematic review of generative AI and academic workload in higher education. Acta Psychol. 2026;268:107353. doi:10.1016/j.actpsy.2026.107353.
- Lyell D, Coiera E. Automation bias and verification complexity: a systematic review. J Am Med Inform Assoc. 2017;24(2):423-431. doi:10.1093/jamia/ocw105.
EWI Works offers many services that can improve your quality of life. We have developed several cost-effective remote services to help you transition to remote work. Find out more about our Online Training, Services, and Resources.
Follow us on Twitter, LinkedIn, and Facebook for more updates and offers, or sign up for our newsletter below.