AI literacy

What it means

AI literacy means knowing enough about AI to use it with good judgment. That includes what a tool can do, where it tends to go wrong, and when a person should step in. Here is an everyday case. An employee asks a chatbot to summarize a contract. An AI-literate employee knows the summary may be wrong, so they check it against the contract before acting on it.

The European Union's AI Act turned the idea into a legal term. It defines AI literacy as "skills, knowledge and understanding that allow providers, deployers and affected persons, taking into account their respective rights and obligations in the context of this Regulation, to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause."¹ In plain terms, it covers the companies that build AI, the organizations that use it, and the people it affects.

Where the word comes from

The earliest use we found is a column titled "What is Needed for AI Literacy?" by KONISHI Yoko, a senior fellow at RIETI. The page is dated December 25, 2015, and January 13, 2016.² For Konishi, AI literacy is a habit of mind, not a technical skill: "AI literacy refers to being conscious about whether there are any standardized/formatted tasks" that could be handed to AI. The column adds, "We should make it a habit to think about what can and cannot be applied to AI."² We have not confirmed that this was the term's first use.

In 2020, Duri Long and Brian Magerko published "What is AI Literacy? Competencies and Design Considerations" at CHI 2020, a computing research conference. The paper says it offers "a concrete definition of AI literacy based on existing research."³ In 2024 the EU AI Act wrote the term into law, and its AI literacy duty took effect on February 2, 2025.⁴

How others define it

• EU AI Act: "skills, knowledge and understanding that allow providers, deployers and affected persons … to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause."¹
• Long and Magerko (2020): "a set of competencies that enables individuals to critically evaluate AI technologies; communicate and collaborate effectively with AI; and use AI as a tool online, at home, and in the workplace."⁵
• The AI Literacy Framework of the European Commission and OECD: "AI literacy empowers students and teachers to understand how AI systems work, critically evaluate their opportunities and risks, and use them responsibly."⁶
• Konishi (2015): "AI literacy refers to being conscious about whether there are any standardized/formatted tasks" that AI could take on.²

Where they agree and differ

The definitions share three ideas: understand what AI is, judge it critically, and use it well. They differ in focus. The EU definition is about roles and duties under the law. Long and Magerko stress working with AI as well as evaluating it. Konishi's column is about deciding which work to hand over, and about being "emotionally prepared to let go of tasks that AI can perform in our place."²

The strongest objection is that the term is too loose to be useful. One 2025 review of research on AI literacy in schools and universities finds that "its definition remains vague" and that "There is little consensus among researchers and practitioners on how to discuss and design AI literacy interventions."⁵ We take that seriously. A vague duty is easy to meet on paper. Our view so far, which is a hypothesis rather than a verdict, is that the term only becomes useful once an organization defines it for its own tools and roles. Evidence that broad, general training changes how people use AI at work would change our mind.

In the news

In 2026 the EU narrowed its AI literacy rule. The AI Act originally required providers and deployers to take measures "to ensure, to their best extent, a sufficient level of AI literacy" among their staff.⁷ An amending law, which the Future of Privacy Forum calls the AI Omnibus, "was published in the Official Journal of the EU on 24 July 2026."⁷ The amended Article 4 now says providers and deployers "shall take measures to support the development of AI literacy," and adds: "This obligation does not require providers or deployers to guarantee any specific level of AI literacy of any individual."⁸ The European Commission sums up the change: "Ensure AI literacy remains an obligation for providers and deployers of AI systems, but no specific – or 'sufficient' - level is mandated."⁴

What it means for your organization

Under the EU rule, the duty still applies. The Commission says "Article 4 of the AI Act does not entail an obligation to measure the knowledge of AI of employees" and "There is no need for a certificate," and that "The supervision and enforcement rules apply from 2 August 2026 onwards."⁴ So the open question is what your organization's training measures actually contain, not whether staff hold a certificate.

Our working view, which we are still testing, is that organizations using [AI agents](/definitions/ai-agent) need one more piece of literacy: knowing whose instructions an agent should follow, and when a person must step in (see [principal](/definitions/principal) and [prompt injection](/definitions/prompt-injection)).

Ask your team: For each AI tool we use, can the people using it say what it gets wrong and when they should check its work?