.png){kind=logo}
AI adoption is changing workplace tasks and increasing concerns about employee skill erosion.
Critical thinking helps employees evaluate, question, and responsibly use AI outputs.
Companies must clarify responsibilities between human-led, AI-assisted, and AI-executed workplace activities.
Artificial intelligence is changing what employees do at work. Still, a new concern is emerging alongside the push for automation: whether workers are getting enough opportunities to practice the human skills needed to oversee AI.
The latest research from the IBM Institute for Business Value, conducted with Oxford Economics, puts critical thinking, decision-making, and evaluating AI outputs at the heart of this discussion. For their study, the researchers surveyed 1,500 CHROs and senior executives involved in workforce strategy decisions, as well as 8,800 employees globally.
The study was carried out in April-June 2026 and covered executives from 21 countries and 23 industries, and employees from 28 countries.
The workplace skills conversation has traditionally focused on teaching employees how to use new technologies. AI is changing that equation.
IBM’s study found that 60% of employees say skill erosion directly affects them, while 46% of executives cite it as a top concern. Critical thinking is among the capabilities employees are most concerned about losing as AI takes on more cognitive work.
This issue differs from a conventional skills gap. A skills gap means employees need to develop capabilities they do not currently have. Skill erosion refers to capabilities employees already possess but may practice less frequently because AI is doing more of the work.
That distinction could matter as businesses increasingly rely on AI for research, analysis, writing, recommendations, and other knowledge-based tasks.
The IBM research found that executives and employees broadly agree that critical thinking and problem framing are important in AI-enabled workplaces.
57% of executives and 49% of employees identified critical thinking and problem framing among the important capabilities required for working with AI.
However, the gap was much larger when it came to assessing AI output.
71% of executives emphasized the ability to oversee and validate AI output, while only 38% of employees considered this aspect crucial. This is a 33-percentage-point gap between executives' expectations and employees' current priorities.
This matters because AI responses are often convincing, even when they contain mistakes. Employees need to discern when an AI response is helpful, when it needs validation, and when they should switch back to human interaction.
Also Read: Why Product Managers Need AI Strategy Skills in 2026
The research also shows how quickly AI is altering everyday responsibilities.
52% of employees said the tasks assigned to them had changed over the previous year as AI reshaped their day-to-day work. Yet only 26% of organizations clearly define work across human-led, AI-assisted, and AI-executed activities.
This creates a practical problem for companies. When responsibilities are unclear, employees may not know who is ultimately accountable for an AI-related decision.
IBM found that 36% of executives said unclear accountability complicates AI deployment. The study also found that organizations that deliberately define how people and AI share work report different outcomes around employee confidence and workflow performance.
Companies are already investing in reskilling. IBM found that 80% of organizations have a reskilling roadmap designed to help employees collaborate with technology. But training employees to use AI does not automatically address the possibility that some existing human capabilities may weaken through reduced practice.
This is where critical thinking skills matter. It is essential not just to know how to initiate or manipulate AI software, but also to challenge its findings, detect assumptions and mistakes, and evaluate whether the answer fits the real-world situation.
Critical thinking skills under the IDC Human Skills Framework for Agentic AI include problem-framing, assumption detection, hallucination detection, trade-offs analysis, and metacognition when working with AI.
Also Read: CHRO Guide to AI Workforce Planning, Skills, Roles for 2027
The research suggests that organizations face a broader challenge than simply introducing AI tools.
IBM found that only 26% of organizations clearly distinguish human-led, AI-assisted, and AI-executed work. Meanwhile, 46% of organizations do not involve the CHRO when defining AI strategy.
That places greater importance on workforce planning. Companies need to decide which decisions should remain human-led, where AI can assist, and where autonomous systems can execute tasks.
HR’s role is also changing. IBM found that 71% of CHROs identify digital and AI fluency as their top personal development priority, while 72% of organizations make limited or no use of AI inside HR itself.
The growing focus on critical thinking does not mean AI has failed to add value. On the contrary, it reflects the growing role of the human element in certain areas.
As AI takes on most cognitive tasks, workers can focus on analyzing results, framing problems, making decisions, handling exceptions, and working through cases when they can't act on automatically generated recommendations.
This implies that for companies, productivity gains cannot be measured simply by the number of activities AI performs; the quality of human involvement also matters.
According to the IBM study, companies that best combined humans and AI enjoyed twice as much revenue growth, workforce benefits, and innovation as other companies.
The AI skill gap is shifting from whether people can use the new technology to whether they can work effectively with it.
Thinking, decision-making, problem framing, and assessing AI outputs are emerging as critical workplace skills, as the onus of decision-making does not disappear when AI does much of the underlying work.
For companies, it is important that AI adoption does not dilute the critical skill of questioning and improving AI outputs. For workers, working with AI without losing independent thinking could become a critical workplace skill.
A good follow-up to this story would explore how susceptible certain jobs are to skill erosion and what human skills employers want in 2027.
1.Why is critical thinking important in AI-driven workplaces?
Critical thinking helps employees evaluate AI outputs, identify errors, question assumptions, and make informed decisions instead of accepting results automatically.
2.How is AI contributing to workplace skill erosion?
Employees may practice certain cognitive skills less when AI increasingly handles research, analysis, writing, recommendations, and other knowledge-based tasks.
3.What skills should employees develop alongside AI literacy?
Employees should strengthen critical thinking, problem framing, judgment, AI-output evaluation, communication, adaptability and decision-making to work effectively alongside AI systems.
4.How can companies prevent AI-related skill erosion?
Companies can provide continuous training, maintain human oversight, encourage independent problem-solving, and clearly define responsibilities across human and AI-driven workflows.
5.Will AI replace critical thinking in the workplace?
AI can support analysis and recommendations, but employees remain responsible for evaluating outputs, understanding context, making judgments, and handling complex decisions.
