AI and talent: Learning to think in the middle

In the 1940s, science‑fiction writer Isaac Asimov imagined intelligent machines that could reason, assist, and work alongside humans. What he called robots, we now call artificial intelligence (AI). What is uncanny is not that he anticipated the technology decades before it arrived, but that he anticipated our behaviour around it. While Asimov foresaw both the sceptics and the enthusiasts, his deeper insight lay elsewhere. Those who succeeded were neither fearful of nor deferential to machines, but learned to think alongside. Eight decades on, that observation reads less like science fiction and more like a blueprint for modern talent strategy.

As organisations deploy generative AI at scale, a familiar pattern emerges. An initial surge of excitement and experimentation, followed by a plateau, and within it a divide. A small cohort integrates AI meaningfully into how they think, prepare, analyse, and decide, while a larger group slips back into old habits, using AI superficially. Some resist AI as unsuitable for nuanced work, while others over‑rely on it for speed and convenience. Both miss the point: one by refusing augmentation, the other by eroding critical thinking.

For leaders, this creates a paradox. AI’s potential is evident, yet its impact varies dramatically across teams and individuals. The constraint is rarely technological but more mindset and behaviour-driven. Where AI adoption is a rollout exercise measured by licences issued, training delivered, and usage tracked, progress stalls. Real progress requires deeper introspection: where does judgement truly add value, which decisions require deeper human engagement, and how must work itself be redesigned rather than simply layered with new tools? These answers shape not only how AI is used, but where it should not be used at all.


This distinction is already visible in different parts of the enterprise. In finance, effective CFOs use AI not as an answer engine but as a sounding board, to surface anomalies, test assumptions, and force sharper questions. In tax functions, companies are using domain‑specific platforms that fit real decision cycles: for example, EY’s AI‑enabled tax solutions embed legislative intelligence, risk surfacing, and scenario modelling directly into workflows. AI-enabled solutions enable tax teams to manage complexity faster while retaining accountability for final positions, freeing professionals to focus on higher-value judgement and advisory work.

At the individual level, AI is compressing learning curves. Employees can now acquire micro‑skills on demand: a finance analyst prototypes code, new solutions are being imagined — all with AI as an active assistant. This form of non-linear learning is beginning to reshape how capability develops inside organisations. This is resulting in agile talent that can increasingly move across roles based on skill adjacencies rather than linear career paths, allowing organisations to redeploy capability faster than traditional hiring cycles permit. With AI mapping skills in real time, leadership teams can reskill existing teams to pursue new opportunities with the question shifting from “Do we have the talent?” to “How quickly can we build it?”

In the process, AI is also beginning to level certain professional advantages. For much of history, access to insight was gated by tenure and proximity. Today, a curious and motivated professional early in their career can develop analytical depth and contextual understanding that once took years to accumulate. Experience, judgement, ethics, and relationships still matter, but the slope of development has changed. Those willing to learn aggressively can now compound faster.

These shifts expose the limits of traditional workforce planning. Role‑based models struggle to keep pace with the speed at which skills evolve. Learning velocity, how quickly people acquire, apply, and adapt new capabilities, is becoming more valuable than credentials or job titles. In parallel, new roles are emerging, from AI governance and human‑machine workflow design to bias oversight. The talent pipelines for these capabilities will require deliberate, sustained building.

Asimov imagined this moment long before the technology arrived, and his insight was always about humans engaging with them. The real leadership call is no longer choosing whether to adopt AI, but deciding where human judgement must remain non‑negotiable. Organisations that get this balance right will compound capability while others merely automate activity. Those who lead the next decade will be the ones who learned — as Asimov implied all along — to think in the middle.

The article is contributed by Giselle H Barboza, Tax Partner, EY India.

Disclaimer: The opinions and views expressed in the article are those of Giselle H Barboza, Tax Partner, EY India, and are provided only for general information purposes. The news and editorial staff of ET had no role in the creation of this article nor vouch for or endorse this content.