Some organisations are pursuing aggressive AI-first strategies, rapidly integrating the technology across workflows, while others take a more measured approach, slowed by
operational or regulatory concerns.

Beyond the corporate world, however, adoption shifts from systemic mandates to personal survival. For independent consultants and small business owners, AI is an empowering asset that slashes overhead and accelerates execution, yet behind this efficiency lies a constant threat of being rendered obsolete by automated competitors.

Employees inside traditional corporate structures view the same tools through a more
complicated lens. For them, efficiency rarely translates to less work or more pay. Instead,
they face immediate anxieties over job security, while grappling with soaring performance
expectations. This pressure is driving the rise of "shadow AI", where employees turn to
unsanctioned tools to keep up.

At this stage of adoption, it is clear the productivity gains are real. AI is helping organisations move faster, automate repetitive tasks, generate content at scale and increase operational efficiency in ways previously impossible.

But this efficiency focus may be masking a blind spot: are we underestimating the human input still required to make those efficiency gains actually work in practice?

To see exactly how much human oversight is still required, one only needs to look at the creative and media industries.

Why Human Input Still Shapes Quality in Creative and Media

Creative sectors are among the clearest demonstrations of AI-driven productivity gains. Scripts, visuals, edits, voiceovers and entire content workflows can now be produced at extraordinary speed and scale, lowering barriers and accelerating experimentation.

As Theo le Roux Preis, Founder of TX VFX and a specialist in visual effects, animation and AIdriven creative workflows, explains, "AI has democratised processes that were previously gated by high barriers to entry. In our business across visual production, optimisation and coding, the biggest gain is high iteration at low cost, especially in creative work, where it enables rapid exploration that was previously too expensive or time-consuming."

But as AI tools rapidly evolve and become more capable, another reality is becoming visible and speaks directly to the broader productivity conversation.

"What's shifted now is that AI is often capable enough to move beyond just being an earlystage tool and is being used for final output too. Used correctly, it's a net positive. I tend to use it upfront in a project rather than at the end, but it raises a tension between efficiency and originality, where human creative input, refinement and craft are still essential to avoid producing work that's derivative and lacking authorship."

That tension increasingly shows up in audience behaviour too. Across digital platforms, there is growing sensitivity to content that feels overly synthetic or emotionally flat. The issue is not necessarily always about AI itself, but the absence of perceivable human authorship behind the work.

A recurring audience sentiment can be summarised simply, "If it feels AI-generated, I disengage".

This exposes a fundamental flaw in the productivity conversation, while automated workflows
can exponentially increase output, human creativity sometimes determines whether an
audience actually connects with or trusts the final result.

The Human Layer Across Industries

We see this same underlying pattern playing out across industries, where AI consistently
improves speed and efficiency, but still depends on human oversight to make outputs reliable and usable in practice.

In healthcare, diagnostic and administrative workflows are accelerated, yet clinicians still carry
responsibility for interpretation and outcomes.

Customer service environments reveal a similar tension. While automated systems improve
response times, companies increasingly face consumer frustration when automation replaces
moments requiring human engagement, empathy, or context.

Legal, financial and professional assurance sectors show the same structural reality. AI
dramatically improves research and analysis efficiency, but accountability and final
interpretation remain firmly with human professionals.

Even in software development, where AI coding tools are now deeply embedded in daily
workflows, experienced engineers remain essential for validation, oversight and maintaining
system integrity.

The broader reality, then, is not whether AI creates productivity gains. It clearly does. The
more nuanced question is whether organisations are sometimes measuring those gains too
narrowly, while underestimating the ongoing human judgement and oversight still required to make AI-generated output truly effective given the current capabilities of these systems.

At this stage, what counts as, "productive" is still being defined in real time, as organisations
recalibrate how AI fits into workflows and decision-making. Alongside this are concerns
around job losses, role redesign and how human capital is valued as tasks are automated.

The question many industries continue to wrestle with is no longer just how AI changes work in the present, but how fundamentally the relationship between human contribution and machine capability may shift over the next five to ten years, as more advanced forms of AI, and potentially AGI, move closer to reality.

For more information, visit www.fourzcommunity.substack.com.

*image courtesy of contributor