The new AI skills gap isn’t technical. It’s human

Organisations are investing in employees’ technical readiness for AI faster than in the human capabilities needed to use it effectively. Workforce data shows that while 50% of employees report receiving technology training in the past three years, only 20% report development in critical thinking. This imbalance creates a workforce-development priority for HR leaders: closing the gap between employees’ access to AI and their capacity to question, interpret, and act effectively on what it produces.

The capabilities this environment requires are not new. Communication, critical thinking, judgement, psychological safety, adaptability and collaboration have always been fundamental to effective work. The speed of AI-enabled work has made weaknesses in those areas more visible and potentially more consequential, given the impact on decision quality, collaboration and organisational risk. AI did not create the human skills gap; it accelerated its consequences.

When speed outpaces scrutiny

AI can accelerate information retrieval, content generation, analysis and routine work. But speed and efficiency create value only when the foundations for sound execution already exist. As AI-generated outputs become more common and appear polished and authoritative, employees’ ability to question assumptions, compare alternatives and evaluate context becomes more important.

Without critical thinking, employees may accept information without adequately evaluating its accuracy, relevance or implications. Without psychological safety, they may recognise a problem but remain reluctant to question an output, process or decision. Without judgement, they may move quickly while still moving in the wrong direction. The risk is not simply that AI can be wrong; it is that errors can now travel faster.

Organisations face a broader challenge in how they use the time AI frees up. When a task that once took an hour can be completed in 20 minutes, employees often fill the time saved with additional tasks. Efficiency can create a cycle in which organisations produce more while having less capacity to evaluate that output. Critical thinking interrupts that cycle. It moves employees beyond completing more tasks faster and toward determining whether the work itself is correct, useful and aligned with the organisation’s direction.

This is where judgement is increasingly valuable. Organisations can access similar software, hardware, data and processes, but similar access does not produce equivalent results. What differs is how people use those resources. Judgement draws on critical thinking, experience, and context to determine what information matters, what needs further scrutiny, and what action should follow. An AI-generated answer can be technically plausible yet inappropriate for a particular customer, employee, market or business decision.

As access to technology expands, differences in outcomes will increasingly reflect how effectively employees interpret information, apply context and exercise judgement. Technology may accelerate the process, but people ultimately remain accountable for the quality of the decision.

AI makes communication more, not less, critical

Communication training often concentrates on managers and leaders, emphasising strategy, change and feedback. But communication is not just a sender capability.

At the individual level, employees interacting with AI need to articulate what they want to accomplish, provide relevant context, and assess the information returned. That requires framing tasks clearly, explaining constraints, asking follow-up questions, and identifying and communicating what remains unknown. At the organisational level, effective communication determines whether people can challenge assumptions, compare perspectives, raise concerns and refine ideas collaboratively.

Effective communication requires a functioning feedback loop. For example, training managers solves only part of the problem if employees are not equally prepared to ask questions, clarify information, challenge conclusions, and contribute expertise. AI increases the importance of those feedback loops because the pace of work leaves less room to correct problems after they spread. Organisations that want to benefit from the collective intelligence of their AI-enabled workforce need communication systems that operate across organisational hierarchies, not just down them.

Psychological safety also plays a role. Employees need both the skill and confidence to question information. Critical thinking and communication provide limited value when organisational culture discourages people from challenging a leader, expert, established process or AI-generated output. Expertise also does not reside exclusively at the top of the hierarchy. Employees closest to the work may hold context that materially changes a decision, making their ability to contribute an essential component of effective AI-enabled work.

Rethinking development for an AI-enabled workforce

HR and learning leaders require a broader view of development. Technical and AI training remain necessary, but the choice is not between technical and human skills. The priority is connecting them.

AI training can deliberately build the human capabilities needed to verify outputs, explain the reasoning behind decisions, identify when additional expertise is needed and raise concerns in psychologically safe settings. Performance expectations, team routines and coaching should reinforce these behaviours rather than treat them as one-time course objectives.

Because these behaviours depend on more than direction from managers and leaders, development must extend across organisational levels. Communication and other interpersonal development should not be reserved for those with direct reports. Employees at every level need to ask better questions, clarify expectations, provide context, and raise concerns as AI becomes more embedded in daily work.

READ MORE: AI risk debate: Steve Boese on what HR leaders can act on while consensus is missing

Development programmes need to bring organisational levels together for training. Although some learning needs to remain role-specific, shared exercises in questioning AI outputs, explaining decisions and responding to dissent can help close existing communication gaps within organisations and strengthen collaboration.

The future of development

As technologies evolve, more durable investment will go to broader human capabilities, including communication, critical thinking, judgement, collaboration, and adaptability. These capabilities are beneficial regardless of the next innovation. Organisations can expand how they evaluate AI’s contribution. Time saved and output produced matter, but so do better decisions, earlier problem identification, reduced rework, knowledge sharing, and improved ways of working.

AI can increase what employees produce. HR’s larger development challenge is ensuring they are equally prepared to determine what deserves to be produced, what deserves further scrutiny and what should happen next. As AI becomes more capable, human responsibility does not disappear. It grows.

About the Author: Robert A. Coleman, Ph.D is the director of research and thought leadership for Dale Carnegie, where he is responsible for ongoing research into current issues facing leaders, employees and organisations worldwide. This article was first published on HR Executive.