We keep hearing that teachers need AI training. But what does that actually mean?
For most institutions, it means a workshop on prompt engineering or a demo of a few AI tools. Check the box, move on. The assumption is that if teachers know how to use the technology, they’re ready to teach with it.
That assumption is wrong. And this study by Bilbao-Eraña and Arroyo-Sagasta (2025) explains why.
What AI Literacy Actually Requires
The authors argue that AI literacy is a foundational competence for teachers, one that extends well beyond technical skills. It includes cognitive, ethical, social, affective, and practical dimensions. In their words: “
“AI literacy goes beyond technical knowledge or coding skills: it includes a combination of cognitive, social, and ethical competencies that enable individuals to make informed decisions about AI in realworld contexts (Lee et al., 2024). In this sense, AI literacy represents not only a response to technological advancement, but also a civic and educational imperative (Long and Magerko, 2020).” (p. 2).
That’s a much broader definition than what most teacher education programs are working with right now. And it’s the right one.
A teacher who knows how to use ChatGPT but can’t think critically about its outputs, can’t navigate the ethical questions it raises, and can’t help students understand how AI systems actually work? That teacher has tool proficiency. They don’t have AI literacy.
The Teacher Preparation Problem
Here’s where it gets uncomfortable. Pre-service teachers are the ones who will be in classrooms for the next 30 years, teaching students who will live and work alongside AI for their entire careers. And yet, most teacher education programs have barely begun to address AI in any meaningful way.
The authors are direct about this: “To fulfill this role, teachers themselves must first be equipped with the necessary knowledge and confidence to address AI in their classroom” (p. 2).
I’ve said this before in previous posts. When I covered the European Commission’s JRC study on AI in secondary schools across five EU countries, one of the strongest findings was that teachers across the board felt underprepared. Professional development on GenAI was scarce or nonexistent.
In Germany, teachers predicted it would take years before AI made its way into training programs. The gap between what students are already doing with AI and what their teachers know how to guide them through is where the real damage happens.
This study confirms the same problem from the other end of the pipeline. If we’re not preparing future teachers during their initial training, we’re just perpetuating the cycle.
What 8 Hours Can (and Can’t) Do
The study tested a short, concept-driven intervention with pre-service teachers. Just 8 hours. No coding. No advanced technical instruction. The focus was on foundational AI concepts, awareness of AI in everyday life, and European AI governance initiatives.
The results? Awareness improved significantly. Participants came away with a better understanding of AI concepts, could recognize AI in tools they use daily, and gained familiarity with policy frameworks they hadn’t heard of before the training.
Attitudes toward AI also improved significantly. Participants became more open, more comfortable, and more likely to see AI as relevant to their professional practice.
That’s meaningful. Eight hours of conceptual instruction moved the needle on both awareness and attitudes. It suggests that a lot of the resistance and anxiety teachers feel around AI comes from unfamiliarity, and that even a relatively brief, well-designed intervention can start to address it.
But here’s what didn’t change: trust.
“The findings from this study indicate that the training program based on AI foundations significantly improved participants’ awareness and attitudes toward AI, while the dimension of trust did not show a statistically significant change” (Bilbao-Eraña & Arroyo-Sagasta, 2025, p. 11).
Trust is different from awareness and different from attitude. You can understand what AI is and feel positive about its potential, and still not trust it. Trust requires deeper engagement with questions of reliability, ethics, governance, and accountability. It develops slowly. It needs sustained, hands-on experience and honest conversations about where AI fails, who it harms, and what safeguards exist.
The authors put it well: “These results underline a key pedagogical implication: AI literacy requires more than conceptual instruction and must articulate cognitive, social and ethical aspects” (p. 13).
Why This Connects to Everything Else
If you’ve been following my recent posts, you’ll see a thread running through all of them. The Shaw and Nave (2026) paper on cognitive surrender showed that students defer to AI outputs without critical evaluation, especially when those outputs sound confident and fluent. The Cheng et al. (2025) study showed that students who ask AI direct, targeted questions perform better, but that skill needs to be taught.
All of those findings point in the same direction: students need guidance. And guidance requires teachers who are prepared.
This study adds the missing piece. You can’t guide students through AI if you haven’t developed your own AI literacy first. And that literacy has to go beyond tool proficiency. Teachers need to understand how AI works, what its limitations are, what ethical questions it raises, and how to create learning environments where students engage with AI critically.
Eight hours is a start. It moves awareness and attitudes. But trust, critical engagement, and the confidence to actually teach with AI in a thoughtful way? That takes a longer, deeper commitment from teacher education programs.
The encouraging takeaway is that even a modest investment in AI literacy training produces measurable results. Imagine what a full, integrated approach across an entire teacher education curriculum could do.
We’re not there yet. But studies like this one show us the path.
Related: AI Research Summary
Reference
Bilbao-Eraña, A., & Arroyo-Sagasta, A. (2025). Fostering AI literacy in pre-service teachers: Impact of a training intervention on awareness, attitude and trust in AI. Frontiers in Education, 10, 1668078. https://doi.org/10.3389/feduc.2025.1668078