Your firm doesn't need coders: the surge in AI fluency 

The most valuable AI skill in your business isn't technical. It's the ability of ordinary employees to work alongside AI well, and the data shows that's where the real shift is happening. 

The skill that's actually rising 

There's a persistent assumption that preparing for AI means hiring engineers. Build a data science team, recruit some machine-learning specialists, and the AI question is handled. The evidence points somewhere quite different. 

According to LinkedIn's Skills on the Rise 2025 analysis, AI Literacy is the single fastest-growing skill globally,  its number-one skill on the rise. Crucially, LinkedIn notes that the people adding "Artificial Intelligence" to their profiles are increasingly non-technical professionals using tools like ChatGPT and Copilot in their everyday work, not engineers learning to build models. The number of LinkedIn members adding AI skills to their profiles has grown roughly twenty-fold globally since 2016, with some markets and sectors far higher. 

This is the surge that matters. Not a wave of new coders, but a far larger wave of ordinary employees, in finance, HR, operations, marketing, client services, becoming fluent in working with AI tools. Microsoft and LinkedIn's 2024 Work Trend Index, based on a survey of 31,000 people across 31 countries, captured the scale of it: a dramatic increase in professionals adding AI skills to their profiles, and most leaders saying they wouldn't hire someone without them. 

The lesson for a mid-sized business is liberating. You almost certainly don't need to build a team of coders to benefit from AI. You need your existing people to become fluent and confident in using it well, and you need that to happen safely. 

Why the future of work is collaborative, not automated 

The most authoritative data on how AI changes work comes from the World Economic Forum's Future of Jobs Report 2025, which surveys around 1,000 large employers worldwide every two years. 

Its central finding on the human-machine relationship is precise and worth quoting in structure. Today, employers estimate that 47% of work tasks are performed mainly by humans, 22% mainly by technology, and 30% by a combination of both. By 2030, those proportions are expected to settle into a near-even split: roughly 33% human, 33% human-machine collaboration, and 34% technology. 

Read that middle column carefully. The human-machine collaboration share, the work done by people and machines together — is the band that's growing into the largest single mode of work. The future the data describes isn't one where machines replace people, or where people carry on untouched. It's one where the dominant pattern is collaboration: humans and AI working on the same tasks, together. 

This reframes the entire AI skills question. If the largest share of future work is collaborative, then the most valuable workforce capability isn't building AI or avoiding AI. It's working with it. The WEF's own data backs this up: the most common workforce strategy employers report is reskilling and upskilling the existing workforce for AI collaboration, cited by 77% of organisations. Hiring people who can work alongside AI (62%) ranks well above hiring people who can build it. 

The skills the WEF identifies as rising fastest alongside AI are revealing too. Analytical thinking, resilience, flexibility, leadership, creative thinking, curiosity. These are human-centric capabilities, the things people bring to the collaboration. AI fluency, in other words, is not the ability to operate a tool. It's the ability to bring human judgement to bear on what the tool produces. 

Why banning AI is the worst possible response 

Faced with the risk side of AI, data leakage, inaccurate outputs, compliance exposure, a tempting response for cautious leadership is to ban it. No personal ChatGPT. No unapproved tools. Lock it down until we understand it. 

The data is unusually clear that this backfires. Research cited across multiple 2025 studies consistently finds that nearly half of employees would continue using personal AI accounts even after an organisational ban. A Software AG study found 46% of shadow-AI users would keep using the tools even if explicitly prohibited. Microsoft's 2025 Work Trend Index found that a large majority of workers have used unapproved AI tools at work, a figure that holds remarkably consistently across industries, company sizes, and seniority levels. This is not a junior-staff problem; security leaders and senior executives are among the most frequent users of unsanctioned tools.

The mechanism is simple. Employees use AI because it makes their work faster and easier, and the tools are free and instantly available. A ban doesn't remove the incentive; it removes the visibility. The work still gets done with AI, it just moves into personal accounts, unmanaged tools, and channels the business cannot see, govern, or protect. IBM's 2025 Cost of a Data Breach Report put a number on the consequence: organisations suffering a breach involving shadow AI faced an average of around $670,000 in additional costs, and the overwhelming majority of AI-related breaches occurred where adequate AI access controls were absent. 

A ban, in short, doesn't exclude AI from the business. It excludes the business from oversight of its own AI use, and it excludes staff from doing openly and safely what they are already doing covertly. The productivity is lost, the risk is retained, and the people who could have been trained are instead driven underground. 

What the firms getting this right actually do 

The organisations capturing value from AI are not the ones with the strictest bans or the biggest engineering teams. They are the ones investing in the fluency of their existing workforce, within a framework that keeps it safe. 

The named examples are instructive. IKEA announced an AI literacy programme aimed at tens of thousands of workers and managers; by August 2024 it had reached around 40,000 employees, and the programme explicitly includes training on responsible AI and AI ethics to reflect the company's values in how the technology is applied. Moderna launched an internal "AI Academy" as early as December 2021, and, importantly, learned that a one-size-fits-all course didn't work, reworking its curriculum into multiple tracks at different skill levels so that basic users and advanced users were each served appropriately. JPMorgan Chase has rolled out AI literacy across its workforce, with specialised tracks for trading, risk, and compliance teams. 

The pattern across these programmes is consistent: broad fluency for everyone, deeper capability where roles require it, and responsible-use training built in from the start rather than bolted on later. Moderna's lesson in particular is one any mid-sized business should absorb, generic AI training quickly loses people; relevance to the actual role is what makes it stick. 

There is a UK-specific urgency to this too. IBM's October 2025 Race for ROI study found that 66% of UK enterprises are already seeing significant AI-driven productivity improvements, but only 38% are prioritising AI upskilling, which IBM identifies as the factor stalling the wider rollout of AI projects. The tools are landing. The fluency to use them well is the bottleneck. 

"Meaningful Human Intervention" isn't just good practice, in the UK it's the law 

There's a crucial point where workforce fluency meets legal obligation, and it's one many businesses miss. 

Under UK GDPR, decisions made about individuals that have legal or similarly significant effects cannot, as a rule, be made by automated processing alone. The framework, Article 22 of the UK GDPR, and now the new Article 22A introduced by the Data (Use and Access) Act 2025, requires meaningful human involvement in such decisions. The Information Commissioner's Office is explicit that this involvement must be genuine: a human cannot simply "rubber-stamp" an AI-generated decision. As the ICO puts it, the review must be carried out by someone with the authority and competence to change the decision, weighing the relevant factors rather than deferring by default to the machine. 

This is where workforce fluency becomes a compliance issue, not just a productivity one. "Meaningful human intervention" only works if the humans doing the intervening actually understand the AI system well enough to challenge it. An employee who can't interrogate an AI output can't meaningfully review it, which means the organisation may be relying on automated decision-making without the safeguards the law requires. Staff fluency is what turns a legal principle into a working control. 

For any UK business using AI in recruitment, lending, pricing, performance management, or any decision that materially affects a person, this is not optional. It is a direct regulatory obligation, and it depends entirely on having people equipped to exercise it. 

Where Turma fits 

This is the gap Turma's learning layer is built to close. 

Turma Passport is the people-compliance and learning layer of the platform. It provides structured AI training for the whole workforce, building the fluency the WEF data shows is the real shift, alongside the evidence that the training has actually happened. It tracks completion, captures attestation, and records the behavioural signals that let leadership know AI is being used safely. Role-relevant learning pathways reflect the Moderna lesson: the right depth for the right person, rather than one generic course that loses half the audience. 

Crucially, Passport doesn't sit in isolation. It feeds into Turma Assured, the governance layer, so that staff fluency and adherence become part of the company's overall AI governance picture, including the evidence needed to demonstrate "meaningful human intervention" is genuinely in place, not just claimed. 

The combined effect is the thing this article has been building towards. Instead of banning AI and losing both the productivity and the visibility, a business can empower its people to collaborate with AI openly and safely, while retaining the meaningful human oversight that UK law requires and good judgement demands. Staff are included rather than excluded. The work comes into the light rather than moving into the shadows. And the human judgement that the WEF, the ICO, and common sense all agree is essential stays firmly in the loop. 

A practical first step 

If you don't currently know how AI-fluent your workforce is, or whether your staff could meaningfully review an AI-driven decision if a regulator asked — the most useful first step is a structured view of where you stand. 

Turma's free Snapshot assessment is built for exactly this. It identifies where your people-side AI readiness is thin, where training would have the most impact, and where the gap between AI use and AI oversight is creating exposure. It takes five to seven minutes. 

Your firm probably doesn't need more coders. It needs its existing people to be fluent, confident, and safe in how they work with AI. That is a learning challenge, not a hiring one, and it is entirely within reach. 

Sources 

• World Economic Forum, The Future of Jobs Report 2025 (survey of ~1,000 large employers worldwide) 

• LinkedIn, Skills on the Rise 2025 (AI Literacy ranked No. 1 fastest-growing skill; ~20-fold global growth in members adding AI skills since 2016) 

• Microsoft & LinkedIn, 2024 Work Trend Index: AI at Work Is Here. Now Comes the Hard Part (survey of 31,000 people across 31 countries) 

• Microsoft & LinkedIn, 2025 Work Trend Index Annual Report 

• IBM, Cost of a Data Breach Report 2025, in partnership with the Ponemon Institute 

• IBM, The Race for ROI study, October 2025 (3,500 business leaders across EMEA, including 500 in the UK) 

• Information Commissioner's Office (ICO), Guidance on AI and Data Protection and guidance on automated decision-making and meaningful human involvement 

• UK GDPR, Article 22; Article 22A as introduced by the Data (Use and Access) Act 2025 

• Software AG shadow AI survey, 2025 (50% of employees using unapproved AI; 46% would continue after a ban) 

• Fortune, Moderna's AI Academy (CHRO Tracey Franklin), March 2024 

• Raconteur / company disclosures, AI training programmes at IKEA, JPMorgan Chase, MasterCard, S&P Global, 2024 

This article is intended as general guidance for business leaders. AI governance, data protection, and employment obligations vary by sector and circumstance, and the Data (Use and Access) Act 2025 provisions are subject to commencement and further ICO guidance. Specific obligations should be confirmed with a qualified solicitor or data protection advisor based on your own facts.