AI literacy: Closing the artificial intelligence skills gap

According to recent research from Sam Manning, a fellow at the Center for Governance of AI, 6.1 million workers in the United States are both exposed to AI and ill-equipped to adapt to the technology. Meanwhile, nearly half of executives recently surveyed by IBM say that their employees lack the AI skills and knowledge necessary to implement AI technologies at scale.

Given AI’s expected impact on industries across the globe, the AI skill gap might have profound results. According to research from the IBM Institute for Business Value, 87% of executives believe that employees are more likely to be augmented than replaced by generative AI.

Their ability to effectively harness the technology’s potential will likely have major ramifications for entire industries: Investment in AI is surging 150% and by 2030 AI is predicted to increase enterprise productivity by 42%. Most organizations—as many as 70%—plan to reinvest those productivity gains back into innovation and growth.

But without strong AI literacy programs, these wide-ranging enterprise transformations might stall. To create the workplace of the future, higher education and enterprises must proactively develop the attitudes and competencies necessary for their employees’ long-term success.

AI literacy is the ability to understand, audit and thoughtfully use AI systems. Rather than a technical skill reserved for engineers, it’s a foundational competency for workers across every function and level of an organization, from entry-level employees to the C-suite.

By understanding AI’s capabilities, limitations and ethical considerations learners gain both practical and conceptual skills. These skills allow the AI-literate to exercise critical thinking in their understanding of AI technologies and their application of AI.

At its most basic, AI literacy means knowing what AI can and can’t do. For instance, understanding that machine learning models identify patterns in data but still require human oversight. More sophisticated forms of AI literacy involve being able to understand bias and risk, along with making informed decisions about how to deploy AI across workflows.

Unlike digital literacy, which applies primarily to knowing how to use software, AI literacy requires a deeper conceptual level.

“As AI handles more routine coding and documentation, professionals are increasingly expected to think holistically,” said Natasha Pillay-Bemath, IBM’s VP of Global Talent Acquisition and Executive Search. Necessarily, their roles are shifting toward “[...] understanding systems end-to-end and validating AI outputs for quality and bias.”

AI literacy isn’t just computer science. It’s knowing how to use AI tools, understanding how they operate, how they reason and which tasks they’re best applied to. It’s also a fundamental change in how work gets done: “AI adoption requires a shift in mindset,” wrote Glenn Dittrich and Kim Morick of IBM Consulting®. “It demands investment not only in technical skills but in capabilities such as empathy, critical thinking and curiosity.”

The pace of AI understanding hasn’t caught up to the dramatic increase in the use of AI. With the release of OpenAI’s ChatGPT in 2022, a wide swath of the public encountered gen AI for the first time. Today, according to Gallup, 12% of all employed adults use AI daily in their jobs, across all sectors. Workers in technology, finance and other high-adoption industries use it more.

But as organizations deploy these tools, from customer service chatbots to generative content tools, many employees lack the foundational knowledge to use them responsibly and effectively.

According to recent research from McKinsey, demand for AI fluency has grown sevenfold in two years, faster than any other skill among job posting in the United States. Meanwhile, the World Economic Forum projects 40% of skills required by the global workforce will change within five years.

The skills gap has real consequences: Employees who don’t understand AI’s limitations might over-trust outputs. Skeptical employees might under-utilize transformative tools. And leadership that fails to understand AI models can’t design implementations that provide real value across an organization. As the technology becomes more sophisticated, these dynamics compound.

What’s more, as businesses implement AI agents and more autonomous software, governance systems become more complex. Intuitive apps, dashboards, no‑code options and natural language queries all increase accessibility but still require AI fluency to operate safely.

And as some AI pilot programs stall and organizations struggle to see immediate results, it has become clear the technology’s potential rests on far more than simple tech implementations. Organizations see more value when they redesign workflows and their employees adapt to new ways of working. All this work requires a nuanced understanding of AI-related workplace tools and skills.

“Entry-level roles are shifting from purely task-driven work to analysis, problem-solving and responsible AI use,” says Natasha Pillay-Bemath, IBM’s VP of Global Talent Acquisition and Executive Search. She notes that learning agility now matters as much as technical skills for new hires.

AI literacy is a cluster of related skills that operate in tangent. Ideally, an interdisciplinary and practical set of competencies tailored to specific outcomes and roles. While the specific track of AI literacy can change depending on the intended level of skill, the World Economic Forum recently introduced a draft curriculum for AI literacy in educational settings. The WEF’s four-pillar AI literacy framework includes:

• Engaging with AI: Knowing the applications of AI, along with evaluating its outputs
• Creating with AI: Understanding how to collaborate with AI tools, keeping in mind ethical considerations
• Managing AI’s actions: Responsibly delegating tasks to AI and ensuring human oversight
• Designing AI solutions: Taking an understanding of how AI works to solve problems in daily life

Some other practical forms of AI literacy might include:

• Understanding how AI works: In a conceptual understanding of AI, a user knows what AI is and how machine learning systems learn from data—as well as the strengths and limitations of various software and algorithms.
• Evaluating AI: In critically evaluating AI, a user can assess AI outputs and decision-making for accuracy, bias and reliability. This process means, for instance, knowing when to verify or when to reject an AI-generated result entirely.
• Applying AI: Practically applying AI involves knowing how to use AI effectively for specific tasks and workflows.

Using ethical, social and governance frameworks for AI: Responsible AI literacy involves understanding the impacts of AI in a broad context: Privacy, fairness, accountability and explainability. Truly AI‑literate users understand that deploying AI is not just a technical decision.

Generative AI literacy is a specific subset of AI literacy focused on large language models (LLMs) and other generative systems. As generative AI tools become more embedded in organizations’ workflows, this dimension has become urgent.

One key aspect is understanding the concept of hallucination and encouraging users to approach AI outputs with a degree of skepticism. Another is prioritizing responsible use. For example, understanding what shouldn’t be shared with external AI systems or recognizing when a task doesn’t benefit from gen AI.

Other important aspects of generative AI literacy include:

• Understanding training data
• Considering IP and copyright issues
• Recognizing model limitations
• Maintaining data privacy awareness

AI literacy is the cornerstone of a robust digital transformation. “Meaningful implementation demands far more than simply activating new technology,” says Kimberly Morick, partner for HR Technology at IBM Consulting. “Success requires a nuanced understanding of how structured and unstructured data complement each other, insight into AI learning mechanisms and strategic human oversight at critical decision points.”

“Without this foundation,” she adds, “organizations risk deploying powerful tools with limited impact.”

Some best practices for building AI literacy in enterprise settings include: 

Effective AI literacy programs start with an honest accounting of where an organization stands. Both in terms of an enterprise’s existing level of AI literacy and the type of literacy it will need in the future. A meaningful baseline assessment can measure literacy across dimensions:

• What tools are being used and how effectively?
• What competencies and conceptual understandings do specific roles need the most?
• What skills already exist in an organization? How can they best be used?

During the baseline assessment, it can be useful to map how roles are expected to change over the near and long term. By planning for job redesign, organizations help ensure that employees will possess the skills they need for the future—and help employees see the real-world benefits of AI literacy.

“By framing every workflow automation as potential for new, creative work, the planning process unlocks potential value,” says Sarah Damenti, Associate Partner for HR and Talent Transformation at IBM. 

In large organizations with many different roles, a single curriculum can sometimes fall flat. One-size-fits all trainings or webinars simultaneously aimed at frontline employees and IT staff can be counter-productive. Successful organizations recognize that different skills are required for different roles and personalize training appropriately. Using AI, organizations can create individualized, role-specific learning paths for employees that consider their existing level of fluency. 

While it’s critical for all employees to possess a fundamental understanding of how AI works, effective enterprise AI literacy programs build learning around the specific tools employees might encounter. This approach bridges the gap between knowing about AI and knowing how to use it effectively in a real workflow. A well-designed AI literacy module walks employees through specific hands-on scenarios rather than training them on abstract concepts or generic tools. 

True AI literacy depends on the ability to critically evaluate AI outputs and potential misinformation in real-world situations. Forward-thinking enterprises teach this skill explicitly, with deliberate exercises and feedback—employees who use AI systems frequently without structured practice in evaluation can become more reliant on these systems over time.

For example, Matt Beane, Associate Professor of Technology Management at UC Santa Barbara, practices an approach that encourages users to reflect on their own choices with AI. As Beane told IBM, he pairs small apprentice teams to work with senior employees who coach small groups through a series of challenges with AI tools. During a debrief session, the groups then answer a series of questions to uncover their unconscious assumptions or find alternative solutions. 

One foundation of an effective AI learning practice is a network of internal champions. People in roles across different functions who are enthusiastic about AI and willing to share their knowledge and experience. Often, enthusiasm for AI literacy trickles down from the top; key stakeholders should be fluent in using AI and explaining how it will bring value to an organization.

When employees are excited about the potential of new tools, they’re more likely to engage deeply with AI literacy programs.

These internal champions can also surface new AI use cases employees are experimenting with or unexpected failures. This feedback loop can be essential to keep policies as well as learning programs connected to the reality of how AI is used across an organization—informing further trainings. 

AI literacy without a strong governance education is incomplete. Effective enterprise AI literacy programs are built alongside a clear governance framework. This approach means establishing explicit policies on acceptable use, teaching users what types of data should be entered into AI systems and how AI-generated outputs should be attributed before use. AI literacy programs should also make governance safeguards visible to employees at all levels, encouraging a culture of compliance.

Learning requires trial-and-error, genuine AI fluency requires cultivating a culture in which iteration is valued. Practically, this approach means dedicating time and resources to experimentation, such as internal hackathons. Morick refers to this mindset as “thinking like a startup.”

“AI is everyone’s initiative,” she says. “Everybody needs to take ownership and say, how am I going to be more effective at my job? How am I going to deliver faster, higher value results to the organization from my work?” By intentionally creating safe spaces for experimentation, organizations can foster long‑term AI literacy and spark a passion for new ideas.

Organizations that treat AI literacy as a permanent organizational priority rather than a one‑time initiative are more likely to succeed. This priority is a pressing imperative when as many as one third of work hours can be automated in the coming years. Organizations that treat AI literacy as a one‑time effort will see their workforce fall behind within months. They also miss an opportunity to develop employee skills and future well-being.

Continuously prioritizing AI learning experiences requires structural commitment and integration into existing talent development processes. It also means thinking differently about skills development over the long term.

For instance, some organizations are doubling down on entry-level hiring, with a focus on fostering a different set of skills. As of 2026, 67% of CEOs believe that AI will increase their entry-level headcount and IBM recently announced it would increase entry-level hires threefold this year. But those roles have been redesigned to prioritize critical analysis and human oversight of AI over more routine manual tasks.

“If we don’t continue to invest in entry-level hires, what happens in 3–5 years?” says Pillay-Bemath. “There’s no pipeline; the well simply dries up.”

Prioritizing AI literacy doesn’t just mean providing professional development to veteran employees. It’s about reimagining how skills are developed across an organization for lasting success.