Teachers are using AI anyway. It’s time to give them a roadmap.

Shadow AI (noun): The widespread phenomenon where educators adopt generative AI tools to manage their workloads (lesson planning, grading, emails) without the knowledge, approval, or guidance of their school administration.

If “Shadow AI” is the symptom of a system catching up, then structured, safe experimentation is the cure. While teachers globally may lack formal guidance1, distinct beacons of hope are emerging that show exactly what happens when you treat educators as partners rather than bystanders.

1. The US model: frameworks over features (AI for Education)

In the United States, the fight against “Shadow AI” is being led by organizations like AI for Education. Co-led by Corey Layne Crouch, a former teacher and school leader, this initiative focuses on moving districts from reactive bans to proactive “AI Literacy.”

Crouch argues that we cannot simply hand teachers a login; we must teach the “knowledge, skills, and mindset” to use it safely. Their solution is the “C Framework”, teaching educators to See AI (understand how it works), use it Safely (privacy/security), Ethically (integrity/bias), and Effectively .

• The Shift: The impact of this training is immediate. Crouch reports that within just 90 minutes, educators can move from uncertainty to confidence, shifting their focus from “how do I use this tool?” to “how can I use this to differentiate for my students?”.

• Policy in Action: This approach is transforming major districts like Chicago Public Schools. By bringing IT teams and academic leaders together, they moved beyond “Shadow AI” to creating a living “AI Guidebook”, now on version five, that constantly evolves as the technology changes.

2. The european model: safety as the gateway (Fobizz)

Across the ocean in Germany, Fobizz tackles the same problem with a focus on compliance. In Europe, where strict GDPR laws make using personal ChatGPT accounts in classrooms legally risky, “Shadow AI” is often a compliance trap.

Fobizz solved this by building a “safe sandbox”: a tool suite where “no personal data is being sent” and the “human in the loop” remains central.

• The evidence: When teachers know they are in a compliant environment, hesitation vanishes. A study conducted with the University of Cologne on Fobizz’s training cohorts found a “drastic change” in educator attitudes.

• The result: As Ksenia Sokolyanskaya of Fobizz notes,

  “Those teachers who are terrified of it... are the ones who’ve had the least exposure to it”.

Once exposure is provided via structured training, the fear dissipates.

The common thread: equity and agency

Despite their different geographies, both initiatives share a critical philosophy: access must be equitable.

• AI for Education provides free prompt libraries and webinars to ensure that high-quality guidance isn’t reserved for wealthy districts.

• Fobizz sells directly to state ministries to ensure teachers don’t have to pay out of pocket, closing the “digital divide”.

The Takeaway

Whether through the policy-driven frameworks of the US or the privacy-first toolsets of Europe, the solution to “Shadow AI” is the same. When provided with a roadmap, educators move quickly from “shadow” users to empowered innovators. As usage solidifies in 2025, the question for school leaders is no longer if they should train their staff, but whether they will provide the safety net that allows that training to take root.

🎧 Listen to our Education Futures episode with Ksenia and Diana from the Fobizz team on Youtube, Spotify, Apple Podcast.

🎧 Listen to our podcast episode with Corey Lane Crouch from AI for education on Youtube, Spotify, Apple Podcast.