What is, what is not, who needs it and why it is difficult to define

It is “the policy of the United States to promote literacy and the competence of AI among the Americans”, reads a executive decree President Donald Trump published on April 23, 2025. The decree, entitled Advancing Artificial Intelligence Education for American Youth, indicates that the advancement of the literacy of AI is now an official national priority.

This raises a series of important questions: what is the literacy of AI, which needs it, and how are you going to build it in a thoughtful and responsible manner?

The implications of the literacy of the AI, or their absence, are of great range. They extend beyond the national ambitions to remain “a world leader in this technological revolution” or even to prepare a “skilled workforce of AI”, as the decree says. Without basic literacy, citizens and consumers are not well equipped to understand the platforms and algorithmic decisions that affect so many areas of their lives: government services, confidentiality, loan, health care, news recommendations and more. And the absence of literacy of the AI ​​risks yielding the important aspects of the future of the company to a handful of multinational companies.

How, then, can institutions help people understand and use – or resist – as individuals, workers, parents, innovators, job seekers, students, employers and citizens? We are a policies scientist and two educational researchers Who study the literacy of AI, and we explore these questions in our research.

What the literacy of AI is and is not

In its foundation, the literacy of AI includes a Mixture of knowledge, skills and attitudes who are of a technical, social and ethical nature. According to a prominent definitionAI literacy refers to “a set of skills that allows individuals to critically assess AI technologies; Communicate and collaborate effectively with AI; And use AI as an online, home and work tool. ”

The literacy of AI is not simply the programming or mechanics of neural networks, and it is certainly not only rapid engineering-that is to say the act of carefully writing invites to chatbots. MoodOr using the AI ​​to write software code, could be fun and important, but restrict the definition of literacy to the last trend or the last need for employers will not cover the long -term bases. And while a single defining of master may not be necessary or even desirable, too much variation makes it difficult to decide on organizational, educational or political strategies.

Who needs AI literacy? Everyone, including employees and students who use it, and citizens struggling with their growing impacts. Each sector and sphere of society is now involved in AI, even if it is not always easy to see for people.

How much literacy that everyone needs and how to get there is a much more difficult issue. Are they enough fast HR training sessions, or do we need to integrate AI In K-12 programs and deliver the university Identification micro-information And practical workshops? There are many things that researchers do not know, which leads to the need to measure the literacy of AI and the effectiveness of different training approaches.

https://www.youtube.com/watch?v=fv2e58rgi9k

Measure the literacy of AI

Although there is an increasing and bipartite consensus on the literacy of AI, there is much less consensus on how to really understand the levels of literacy of people. Researchers focused on different aspects, such as technical or ethical skills, or on different populations – for example, business leaders and students – or even on sub -domains such as generative AI.

A recent review study has identified more than dozen questionnaires designed to measure the literacy of AIThe vast majority of which was based on self -detached answers to questions and statements such as “I am confident to use AI”. There is also a lack of tests to see if these questionnaires work well for people from different cultural backgrounds.

In addition, the rise of the generator has exposed Gaps and challenges: Is it possible to create a stable means of measuring the literacy of AI when AI is itself so dynamic?

In our research collaboration, we have tried to help solve some of these problems. In particular, we focused on creating objective assessments of knowledge, such as multiple choice surveys tested with in -depth statistical analyzes to ensure that they Precisely measure the literacy of AI. So far we have tested a multiple choice survey in the United States, the United Kingdom and Germany and have noted that it works in a coherent and equitable manner through These three countries.

There is much more work to do to create reliable and achievable test approaches. But in the future, it is not enough to ask people to self-assess their literacy of AI different groups of people are and the supports they need.

Approaches to build the literacy of AI

Governments, universities and industry are trying to advance the literacy of AI.

Finland launched the Elements of the AI ​​series In 2018 in the hope of educating his general public on AI. Estonia has a jump The initiative is associated with Anthropic and Openai to give access to AI tools for tens of thousands of students and thousands of teachers. And China is now requiring at least eight hours IA education every year from primary school, which goes beyond the new American decree. At the university level, Purdue University and the University of Pennsylvania have launched new masters in AI programs, targeting future IA leaders.

Despite these efforts, these initiatives are faced with an unclear and evolving understanding of the literacy of AI. They are also faced with challenges to the measure of minimum efficiency and knowledge on educational approaches that actually work. And there are long -standing problems with regard to equity – for example, to reach schools, communities, population segments and companies that are stretched or subdress.

The next movements on the literacy of AI

Based on our research, our experience as educators and collaboration with decision -makers and technological companies, we believe that a few steps could be cautious.

The construction of literacy of the AI ​​begins by recognizing that it is not only a question of technology: people must also grasp the social and ethical sides of technology. To see if we arrive there, researchers and educators, must use clear and reliable tests which follow the progress of different age groups and communities. Universities and companies can first try new teaching ideas, then share what works through an independent center. Educators, on the other hand, need appropriate training and resources, not just additional programs, to introduce AI to class. And because The opportunity does not spread uniformlyPartnerships that reach schools and under-strength districts are essential so that everyone can benefit from them.

Above all, the realization of the widely widespread AI literacy can be even more difficult than creating digital literacy and the media, so getting there will require serious investments – not cuts – for education and research.

There is a large consensus make your future. As with AI itself, we think it is important to carefully approach the literacy of AI, avoiding media threshing or too technical orientation. The right approach can prepare students to become “active and responsible participants in the labor market for the future” and allow Americans to “prosper in an increasingly digital society”, as the AI Literacy Executive Decree Call.

The conversation will organize a free webinar on the practical and safe use of AI with our technological publisher and an expert in AI on June 24 at 2 p.m. he / 11 a.m. PT. Register to answer your questions.