Why AI Literacy Is Now a Baseline Skill, Not a Nice-to-Have

Most companies still treat AI training as optional. Something for the tech team, or a one-off session after a new tool gets rolled out. The data from 2026 suggests that's a mistake, and a costly one.

Everyone is using it. Almost no one has been properly trained.

According to Study.com's 2026 State of AI Jobs and Skills report, roughly nine in ten employees now use AI at work at least occasionally, making adoption effectively universal across most organisations. Yet the same report found that 35% of employees have received no AI training of any kind, and among those who did receive training, only 18% felt it actually prepared them to work independently. In other words: adoption has outpaced preparation by a wide margin, and most employees are learning through unstructured trial and error rather than any deliberate skill-building.

This isn't a story about resistance or incapability. The Study.com researchers were explicit that the core problem is structural — either training doesn't exist, or what does exist isn't aligned with how people actually use AI in their specific roles.

The gap isn't technical — it's applied fluency

DataCamp's 2026 State of Data & AI Literacy report, based on a survey of more than 500 enterprise leaders across the US and UK, found that 88% consider basic data literacy essential for day-to-day work, and 72% say the same about AI literacy specifically. Despite this, around 60% of those same leaders admit to an active data or AI skills gap inside their organisation. Crucially, the researchers found this gap isn't about advanced technical or engineering expertise. It's about foundational, applied AI literacy at the level of ordinary daily tasks.

Separately, Skillsoft's 2026 workforce research found that while 86% of individual contributors and 93% of managers use AI at work, only 24% of individual contributors and 77% of managers strongly agree their organisation has actually prepared them to use it effectively. That's a wide confidence gap between the people doing the day-to-day work and the leaders who assume they're covered.

Why generic "AI 101" sessions tend to fall flat

DataCamp's research points to a specific reason a lot of current AI training underperforms: video-based courses and generic blended learning remain the most common format (used by 40% of organisations), but a meaningful share of leaders say this format makes it difficult for employees to apply what they've learned to real work. Roughly three in five leaders report frustration with third-party training providers on this exact point — learning paths that aren't tailored to specific roles, and employees left unsure where to even start.

Study.com's researchers reached a similar conclusion from a different angle: the employees in the most AI-exposed roles are often given training that's generic rather than tied to the specific tasks AI is already doing in their actual jobs. Their recommendation is that effective AI training needs to be role-specific, built around real examples from an employee's actual responsibilities, with structured practice in evaluating AI output for accuracy — not just passive exposure to what the tools can do.

What good AI literacy looks like in practice

Based on this research, a handful of practical habits separate people who use AI well from those who don't:

• Prompting with intent — being specific about task, audience, and format, rather than issuing vague requests and hoping for the best
• Evaluating output critically — treating AI-generated content as a first draft to be checked, not a finished answer to be trusted outright. This is consistently the skill gap employees report most, according to Study.com's findings
• Understanding data boundaries — knowing what information is safe to input into an AI tool, and where sensitive or confidential data should never go
• Role-specific application — connecting AI use directly to the actual tasks in someone's job, rather than generic capability awareness

These habits are the focus of our Practical AI for Business Teams course, where participants build AI workflows for tasks from their own roles and practise checking the output.

DataCamp's research offers a useful real-world example of what happens when this is done properly: at Rolls-Royce, structured, role-specific upskilling in tools like Python and Power BI for both engineers and non-technical staff reportedly increased data-handling speed dramatically — in some cases by a factor of 100 — because the training was tied to real applied tasks rather than generic awareness-building.

The practical takeaway

Treating AI literacy as a baseline skill — comparable to email or spreadsheets a decade ago — rather than a specialist topic changes how a business should think about training it. It isn't a single workshop to check off a compliance list. Based on the evidence, it's a foundational capability that needs to be built deliberately, tied to real role-specific tasks, with structured opportunities to practise evaluating AI output — not just watching a video about what AI can do.

If you're planning AI training for your team, take our training needs quiz to identify where to start, or book a free consultation to talk it through.