What can we all learn from Korea?

As part of a recent “Leading AI in Schools” day, hosted by Seoul International School and Faria Learn, I had the delight of facilitation a panel of educators working in Korea on the impact of the AI legislation in schools.

The panel discussed the practical implementation of AI in schools, focusing on privacy, data protection, and ethical considerations.

The Korean Basic AI Act, alongside PIPA (Korea’s GDPR-style data protection law), treats education as a high‑risk domain and puts strong emphasis on privacy, data protection, and human oversight. For schools, this means you must be very careful about what student and staff data goes into AI systems, prefer vetted/enterprise tools (e.g. Gemini in a protected Google Workspace environment) over open consumer tools, and avoid sending sensitive data (like passports, photos, grades, or full student records) into unvetted services. The law explicitly requires that a human stays “in the loop” for high‑impact uses such as assessment: if AI is used to grade or help decide outcomes, a human must be able to review, verify, and explain the AI’s decision to a student who asks, “Why did I get this grade?”

As an observer, I was struck by how similar it is in tone and intent to the European Union AI Act, of which I have written before.

Practically, the law pushes schools toward **transparency and documentation rather than banning AI outright**.

Schools need to:

1. clearly disclose when and how AI is used (e.g. in assessments, reports, communications, or learning platforms);
2. document AI use in policies and, ideally, a use‑case/risk registry;
3. ensure staff have enough AI literacy to understand both the tools and their legal/ethical obligations; and
4. avoid “shadow AI” by giving teachers safe, approved options.

The cultural and legal reality in Korea is that there is real liability for individuals and institutions, so leaders and tech directors are trying to build “a sandbox” where teachers know what’s allowed, can use AI to save time and personalize learning, but are protected by clear policies, proper consent, and strong data‑handling practices.

Marc highlighted the importance of adhering to Korean laws like the PIPA Act and the Basic AI Act, emphasizing the need for transparency and human oversight in AI-driven assessments. He spoke to the additional burden experienced by Tech Leaders as the “No police”:

“For myself, I focus a lot on just trying to navigate that world. When people ask me, Can I use this AI, platform or this website, or anything to do with anything Tech, I have to go look at the laws, and that dictates a lot of my time.”

Nick stressed the need for change management and value-add for teachers. He shared his approach to keeping a human-centred value in the midst of seemingly impossible change:

“Change doesn’t have to be scary or transformative. I think if you look at the Diary of CEO Stephen Bartlett talks about 1% better every day. If schools were to look at this and just focus on a one degree shift, those incremental changes over time compound.”

Dion underscored the importance of aligning AI use with a school’s mission and core values.

“It needs to be about if you really value that human centeredness and that connection with the students, you need to make sure you mandate that it’s reinvested, and it needs to come from the top down to say we’re not going to allow that time to be used elsewhere.”

This really resonated with our audience as we see all too easily how efficiencies gained by AI can quickly be eaten up by the never-ending drive for productivity.

Finally, I closed the session with a difficult, possibly unfair question – given that we had an all male panel, why is it that the role of Tech Director either fails to attract or actively excludes women for the most part? I observed that more than 95% of the tech directors I meet are male, and that I often have to work harder to be taken seriously in technical conversations.

I noted a recurring pattern: some tech leaders test me carefully with niche, highly technical follow‑ups to check if you “really” know what you’re talking about – something you don’t experience in the same way with female leaders. This led us into a broader reflection that the bias is often subtle, systemic and internalised rather than overtly hostile.

The panel’s responses added several angles:

• Demographics & pipeline: One leader shared that when they advertised edtech coaching roles (framed heavily as instructional, not just “tech”), ~95% of applicants were men, making balanced hiring difficult in practice.
• Framing of roles: Nick highlighted how “instructional coach” language, rather than “tech” branding, helped attract and normalize women in adjacent coaching roles, suggesting naming/framing matters.
• Early socialisation: Another participant noted that girls often turn away from tech around middle school, so even though many of the strongest technology students they’d taught were girls, this doesn’t translate into equal representation later.

Overall, the exchange suggested that a mix of cultural expectations, job framing, subtle credibility tests, and early pipeline leakage all contribute to the scarcity of women in tech directorships, even when girls and women demonstrate strong aptitude in practice.

I was very grateful to the panel for their expertise, their candour and their willingness to share their knowledge with the room, go raibh míle maith agaibh go léir.

My thanks to the participants for the lively discussion and to Faria Learn for supporting and Seoul International School for the warmest of welcomes.

Looking forward to the next one already!