Entry-level job postings in the most AI-exposed fields grew 35% between 2019 and 2025, according to PwC’s 2026 Global AI Jobs Barometer, which analyzed over one billion job ads across six continents. The roles that shrank – by 10% – were traditional entry-level positions built around routine execution. That divergence tells you something important about where the labor market is actually heading, and it has very little to do with who can prompt an AI tool most efficiently.
The assumption running through most corporate training budgets right now is that AI literacy is the skill gap to close. Load up on workshops about large language models, get everyone comfortable with generative tools, and the workforce will be ready. The numbers suggest that assumption is only half right – and possibly the less important half. What’s moving faster than AI adoption is demand for the abilities that AI still can’t do: reading a room, making a call under pressure, holding a relationship together when things go sideways.
Maria Flynn, president and CEO of Jobs for the Future, calls these “durable skills” – capabilities that hold their value across economic shifts, not just the current one. The more AI absorbs routine cognitive work, the more valuable those durable human abilities become. Five of them are already pulling ahead.
Emotional Intelligence: The Human Skills AI Can’t Fake
Research from TalentSmart found that emotional intelligence is the strongest predictor of workplace performance, accounting for 58% of success across all job types. That finding predates the current AI wave – but it’s more relevant now than when it was first published.
Emotional intelligence (EQ) is the ability to recognize, understand, and manage your own emotions and those of the people around you. It’s what lets a manager deliver hard feedback without destroying morale, or a salesperson sense that a deal is about to fall apart before the client says a word. AI can analyze sentiment in text. It cannot feel the shift in energy when a conversation turns.
Roles using generative AI increasingly require greater emotional intelligence, creativity, and ethical reasoning than comparable roles that don’t use AI. EQ is now listed as a requirement in the same job posts asking for AI fluency. The two are being demanded together.
The practical implication: actively practice reading emotional cues in real interactions. Ask for feedback after difficult conversations. Reflect on moments when you misread someone’s reaction and why. EQ can be built deliberately – it responds to practice in ways that AI capability does not.
Relationship Building: The Competitive Edge No Human Skills Algorithm Has
As AI automates more technical work, competitive advantage increasingly depends on cultivating human connection and the kind of empathy that technology cannot replicate. This is a strategic reality, not a philosophical one.
When every company in an industry has access to the same AI tools – the same models, the same data processing, the same outputs – differentiation collapses at the technical layer. Heather Stefanski, McKinsey’s chief learning and development officer, raised this competitive risk directly in an HR Brew interview: when every firm uses the same AI to arrive at the same answers, the judgment to know which answer fits this customer, this conversation, and this moment comes only from human experience. That judgment lives inside relationships.
The PwC 2026 Barometer identifies a two-track labor market, with “professionalised” roles – those where AI automates routine tasks so that human judgment and expertise are emphasized – growing faster across both headcount and wages. Companies taking a human-AI augmentation approach, using AI to support rather than replace human judgment, are seeing significantly stronger revenue and productivity growth than those that optimized only for technical efficiency.
Relationships require time and presence. The actionable move is simple but underused: schedule regular one-on-one conversations with colleagues and clients that have no agenda beyond connection. Show up physically when it matters. Remember personal details. These behaviors compound over years in ways that no tool can replicate.
Critical Thinking: Knowing When the AI Is Wrong
AI is very good at producing confident-sounding answers. Where it falls short is in flagging when those answers are wrong, biased, or built on incomplete data. That gap is where critical thinking becomes essential – not as an abstract intellectual virtue, but as a practical error-catching function.
Research published in the Journal of Intelligence by Daniela Dumitru and Diane F. Halpern explored AI’s impact on the job market and found that solely relying on AI systems can lead to errors and misjudgments, emphasizing the need for human oversight – particularly skills like critical thinking, problem-solving, empathy, and ethics that machines cannot replicate with the same agility.
Skills needed for the most AI-exposed jobs are changing more than twice as fast as for the least AI-exposed jobs, according to PwC’s 2026 Global AI Jobs Barometer. Among those fast-changing requirements, critical thinking keeps rising. Employers need people who can interrogate outputs, challenge assumptions, and decide when a model’s recommendation shouldn’t be followed.
The practical habit: when you receive an AI-generated output, before accepting it, ask three questions – what assumption is this built on, what would have to be true for this to be wrong, and what does my own knowledge of this context suggest? That friction is the job.
Ethical Judgment: The Human Skill That Keeps Companies Out of Trouble
Entry-level roles with the highest AI exposure increasingly require capabilities traditionally associated with far more experienced workers. Senior-level competencies like leadership, mentoring, and process management now account for 52% of new skills required for highly AI-exposed entry-level positions, compared with just 7% for the lowest AI-exposure roles.
Buried in that shift is something that doesn’t show up in most AI training programs: ethical judgment. AI systems produce outputs at scale and at speed. Every organization deploying them needs humans who can evaluate whether a particular output should be used, whether a data set should be applied to a particular decision, and whether a fully automated process crosses a line it shouldn’t. That evaluation cannot itself be automated without creating a circular problem.
PwC’s Barometer identifies “professionalised” roles – those in which AI automates routine tasks so that human judgment and expertise are emphasized – growing faster across both headcount and wages than roles where AI simply makes the work easier for non-experts. Ethical reasoning is one of the core human-intensive skills driving that growth. Professionalised roles are seeing twice the growth in available jobs and 42% faster salary growth than their democratised counterparts.
Building ethical judgment means engaging seriously with the real trade-offs in your field – not generic ethics courses, but the specific decisions your organization makes about data, automation, and accountability, and what standards those decisions are held to.
Navigating Ambiguity: Comfort With Uncertainty as a Career Asset
The fifth skill is the one most corporate training programs haven’t figured out how to teach. Ambiguity – situations where the data is incomplete, the right answer isn’t obvious, and the cost of waiting for certainty is too high – is exactly where AI struggles most and humans are most needed.
At the entry level, AI appears to be increasing demand for more senior skills from junior workers. Based on 2.4 million entry-level jobs analyzed in the US, entry-level roles most exposed to AI are now seven times more likely to require traditionally senior-level human-intensive skills like leadership, creativity, or face-to-face interactions – and job openings for these roles have grown 35% since 2019, while other entry-level roles shrank 10%.
Leadership and creative judgment – two of the most prominent skills in that growing cohort – are fundamentally exercises in navigating ambiguity. They require making a call when the information is messy, reading a situation that doesn’t have a clean precedent, and committing to a direction without the luxury of certainty. AI tools can provide data to inform those calls. They cannot make them.
A concrete illustration from healthcare: AI automates routine tasks like documentation and administrative duties, giving nurses more time for direct patient care. As nurses manage AI-driven analytics and patient-care technologies, their role becomes even more crucial – with AI supporting decision-making by analyzing vast datasets while nurses focus on clinical judgment and patient outcomes. The administrative burden moves to the machine. The human judgment – reading a patient’s distress, sensing that something is wrong before the numbers confirm it – stays with the nurse. That division of labor is the model playing out across industries.
Read More: AI Could Impact Over 50% of U.S. Jobs, New Analysis Finds
What This Means for You
McKinsey’s 2025 Superagency report found that 92% of companies plan to increase AI investments over the next three years – but only 1% of leaders consider their organization truly ready. That gap exists partly because most readiness frameworks focus on tool adoption and ignore the human capabilities that determine how well those tools actually get used.
Nearly two in three enterprise leaders report a data or AI skills gap in their organization. Organizations with mature AI literacy programs that integrate human skill development alongside technical training see significantly stronger returns on their AI investments than those focused on tools alone.
The workers who will hold the most durable positions are those who invest in both: enough AI fluency to use the tools effectively, and enough emotional intelligence, relational depth, critical thinking, ethical grounding, and tolerance for ambiguity to contribute what the tools genuinely cannot. The skills that AI keeps amplifying the demand for – judgment, creativity, leadership, ethical reasoning – are the same ones that have always separated effective professionals from merely capable ones.
Start with whichever of these five feels most underdeveloped. Emotional intelligence responds to feedback and reflection. Relationship-building responds to consistent presence and attention. Critical thinking sharpens with deliberate practice questioning outputs. Ethical judgment develops through real engagement with your field’s specific trade-offs. Comfort with ambiguity grows through taking on decisions that make you uncomfortable before you feel ready. None of these require a course. All of them require repetition.
AI Disclaimer: This article was created with the assistance of AI tools and reviewed by a human editor.