AI literacy: Closing the artificial intelligence skills gap

Maintenance teams work with drones to monitor bridge conditions in Denmark. Utility workers use artificial intelligence (AI) software to dispatch field staff to clear storm grates in Australia. And auto production employees deploy computer vision to detect manufacturing defects at Ford assembly plants.¹

Each of these examples is a case study of how AI can augment human labor in a sector or specialty. However, effective use of AI in business and industry entails more than companies integrating AI into their computing systems.

It also requires that employees ranging from executives to field workers can use AI tools successfully.

That’s no small task: nearly half of executives recently surveyed by IBM say that their people lack the AI skills and knowledge necessary to implement and scale artificial intelligence technologies.

Fortunately, there’s a solution to this skills gap: fostering AI literacy.

One definition of AI literacy is the ability to comprehend various aspects of artificial intelligence—including its capabilities, limitations and ethical considerations—and to use it for practical purposes. It might entail learners exercising critical thinking in their understanding of AI technologies and the applications of AI.

Being AI literate “requires not just learning but learning to learn—asking the right questions to comprehend how AI systems work,” writes Ignacio Cruz, an expert in communications and emerging technologies.

Cruz adds that AI literacy goals “can span a continuum” ranging from a basic understanding of AI concepts to more sophisticated abilities, such as the ability to evaluate AI risks in automated decision-making.²

Today’s conventional wisdom indicates that AI literacy skills are important not just for employees, but for individuals as they experience the real-world impact of AI.

After all, many people’s daily lives now include an array of AI-related activities, such as reading AI-curated news headlines, interacting with customer service chatbots, riding in AI-guided vehicles, and using consumer-facing generative AI (gen AI) apps including ChatGPT.

“[W]hen people know about artificial intelligence, they’re able to make better decisions for themselves and for their communities,” Charlotte Dungan, COO of The AI Education Project, said in an interview with Harvard Ed. Magazine. “They don’t have to be a programmer to benefit from learning about AI and I think everyone deserves access to that information.”³

While the understanding of what constitutes AI literacy evolves with the evolution of AI itself, there are several foundational principles established by AI researchers and educators in recent years.

One guiding force has been a group of educators and technologists known as the “AI for K12” working group. The group, which is a joint project of the Association for the Advancement of Artificial Intelligence and the Computer Science Teachers Association (CSTA), devised a set of 5 “big ideas” to guide educational standards for AI literacy:

• “Computers perceive the world using sensors.”
  
  
• “[AI] agents maintain models/representations of the world and use them for reasoning.”
  
  
• “Computers can learn from data.”
  
  
• “Making agents interact comfortably with humans is a substantial challenge for AI developers.”
  
  
• “AI applications can impact society in both positive and negative ways.”

Building upon these principles, researchers Duri Long and Brian Magerko at the Georgia Institute of Technology created an AI literacy framework with a set of more than a dozen competencies.

The researchers based their findings on a literature review of 150 papers on AI, such as journal articles and conference papers. The competencies they identified include:

• Distinguishing between technologies that do and don’t use artificial intelligence.
  
  
• Identifying the various technologies that use AI, including technologies incorporating machine learning (ML) and robotics.
  
  
• Identifying the types of problem-solving at which AI excels and what types of problems are more challenging, then using this information to decide when to deploy AI versus when to use human skills.
  
  
• Recognizing the role humans play in programming, choosing AI models and fine-tuning AI systems.
  
  
• Understanding data literacy concepts, including recognizing when personal data is used to train machine learning algorithms.
  
  
• Understanding that some AI systems are able to physically act in the real world.
  
  
• Identifying key AI ethics issues, such as data privacy, explainability, misinformation, bias, transparency and accountability.

The researchers also noted that digital literacy is a prerequisite for AI literacy because “individuals need to understand how to use computers to make sense of AI.”⁴

With the increasing prevalence of generative AI-based applications and tools, some experts have placed a particular emphasis on generative artificial intelligence literacy.

According to Cornell University’s Center for Teaching and Innovation, generative AI literacy includes skills in recognizing when generative AI is being used, assessing the reliability and validity of generative AI outputs and, as with the Georgia Institute framework, identifying AI ethics issues.⁵

Generative AI literacy is especially salient in the realm of higher education, where educators worry that generative AI use might help students “circumvent” learning experiences.⁶

Large language models (LLMs) “might help a learner write a paper or a report, but they cannot teach the learner how to conduct original research, synthesize information from multiple sources, formulate arguments, express opinions or cite sources properly,” the Cornell center advised. “Thus, the need for AI literacy is essential for students and faculty alike.”⁷

So what does it take to achieve AI literacy? Some educators today say that AI literacy education should begin as early as elementary school, while colleges and corporations are offering AI courses and professional development programs to get today’s adults up to speed.

MIT partnered with a STEM education nonprofit organization to develop short-form AI literacy curricula for educators who teach students from ages 5 to 18.

The program called Day of AI introduces AI concepts and vocabulary, encourages hands-on activities that demonstrate how AI works, explores the potential benefits and harms of AI, and stresses the need for responsible AI. Day of AI materials could be downloaded, used and redistributed under a Creative Commons license.

After the curricula were implemented in various schools in 2022 and 2023, researchers surveyed participating teachers and found promising results.

“After only a few hours of engagement with the content, teachers reported that they and their students learned AI concepts, how AI works and is currently being used, and about potential benefits and harms to society,” Cynthia Breazeal, of MIT, and Fiona Hollands, of the research organization Ed Researcher, wrote in an article documenting the initiative. “Learning more about AI increased their levels of optimism about the potential benefits of AI to society and about their own abilities to contribute to shaping the future of AI.” ⁸

Not surprisingly, AI has become a popular area of study for college students. One 2024 global survey found that more than 70% of students believe that universities should offer more courses in AI literacy and training on the effective use of AI tools.⁹

Universities are working to respond to the demand, with some going beyond offering AI classes to awarding undergraduate and graduate degrees focused on AI.¹⁰

Some university-sponsored AI courses are available online for free to the public, including offerings by Harvard University, the University of Pennsylvania and MIT. MIT, through its Responsible AI for Social Empowerment and Education (RAISE) initiative, has a collaborative partnership with Google to provide free AI courses specifically designed for educators.¹¹

Like universities, companies are under increasing pressure to offer AI training. However, this pressure comes not just from employee demand but from a stark reality: a looming AI skills shortage, especially in the case of generative AI.

“A gen-AI-based talent transformation isn’t something companies can simply hire their way out of, as it affects the entire organization and its ways of working,” researchers from McKinsey concluded in a 2024 survey report. The researchers found that early adopters of generative AI technology focus heavily on upskilling and reskilling.¹²

According to a 2024 report from the IBM Institute for Business Value, CEOs say 35% of their workforce will require retraining and reskilling over the next three years—up from just 6% in 2021.

Companies aren’t stopping at offering AI training to just their own employees. Corporations such as Amazon, Microsoft, Intel and IBM are offering free AI courses and learning resources online.

For example, IBM offers SkillsBuild, a program of online courses that award learners digital credentials. Popular courses include “Introduction to artificial intelligence” and “Mastering the art of prompt writing,” and IBM is committed to training 2 million learners in AI by 2026. 

IBM also offers a video series called AI Academy, with episodes featuring subjects such as matching AI models to use cases, AI readiness in the hybrid cloud and an “AI in Action” podcast.

Learn about IBM’s commitment to train 2 million learners in AI

Justina Nixon-Saintil, VP and Chief Impact Officer of IBM Corporate Social Responsibility, predicts that lifelong learning in AI and other technical subjects will become the “new normal.”

“The immediate need is for AI skills, but there will soon be a growing demand for quantum skills, alongside an enduring demand for cybersecurity skills,” Nixo-Saintil explained recently. "Lifelong learning is essential for people who want to remain competitive in the job market.”