[Policy Recommendations]: XR2Learn Policies in Focus - Let’s Discuss

[xr2learn-admin]

Based on policy work led by our project coordinator within XR2Learn, we’d like to open a discussion around a key recommendation here and want to hear from you!

XR in education: what’s really slowing adoption?

With the green and digital transitions accelerating (think Digital Twins, AI, automation), industry is evolving faster than ever.

But education and training? Not quite keeping up.

This creates a key challenge:
how do we equip the workforce with the right skills, at the same speed as industrial change, while keeping a human-centric approach?

XR (VR, AR, MR) is often seen as a powerful solution—enabling immersive and adaptive learning.
Yet, its adoption in education is still limited.

One major reason:
creating XR content is still too complex and inaccessible for most actors.

---

Policy Recommendation #1

A proposal emerging from XR2Learn is to make open-access, low-code / no-code XR tools a requirement in publicly funded projects (e.g. Horizon Europe, Digital Europe, ESF+).

Not just funding content, but funding tools that others can reuse.

Why?
Because today many results remain proprietary and hard to scale.

XR2Learn already showed a different approach:
the INTERACT no-code tool enables even small teams to build advanced XR training scenarios—without specialised programming skills.

The takeaway:
if tools are accessible and open, XR creation becomes scalable across the ecosystem—not just limited to a few players.

---

What do you think?

• Would this kind of requirement actually accelerate XR adoption?
• Are SMEs and training providers ready to use these tools?
• What could hold this back?

This article is based on scientific and technical results from the XR2Learn project , funded by the European Commission under Horizon Europe (Grant Agreement No. 101070300).
The policy recommendations presented here were developed under the guidance of the XR2Learn Consortium Coordinator, Prof. Ioannis Chatzigiannakis (Sapienza University of Rome and CNIT, Italy).

[mari.conte]

Very interesting recommendation. Ensuring that publicly funded projects deliver open and reusable tools, not only content, could be a strong lever to improve scalability and sustainability of results across the ecosystem. XR2Learn clearly demonstrates the potential of no-code approaches to democratise XR development.

[filisia]

Coming from another emerging technology, namely blockchain, my experience suggests that one of the biggest barriers that slow down adoption of such technologies is the technical knowledge gap. That same applies to eXtended Reality. A limited understanding of the technology and how to use it creates mistrust and frustration, ultimately reducing user acceptance. This has been confirmed by multiple studies, which show that although the benefits of the techology are generally recognised, a lack of experience and limited access to resources remain among the main factors discouraging adoption on the individual level.(https://www.researchgate.net/publication/388670237_Barriers_to_the_Adoption_of_Augmented_Reality_Technologies_for_Education_and_Training_in_the_Built_Environment_A_Developing_Country_Context )

Making the technology more accessible and bridging these existing gaps, whether through education and training or through software solutions such as INTERACT, will significantly encourage more people to experiment with and adopt XR technology.

At the industrial/SME level however, I identify two additional barriers to adoption, specifically organizational change, and the shortage of experienced talent. Both of these stem from a more fundamental issue: lack of awareness. This is precisely why paradigms such as XR2Learn are so important. By raising awareness of the technology, demonstrating its benefits across different domains, and, most importantly, fostering a community for support and knowledge transfer, such initiatives can play a key role in accelerating adoption.

[angelica.marsico]

I completely agree: one of the main barriers to XR adoption is the technological gap and limited access to development tools. Thank you for sharing your insights about no-code tools like INTERACT help lower the technical barrier, enabling even small teams and SMEs to create immersive experiences without advanced programming skills.

From an educational perspective, XR is not just a delivery tool, it’s a powerful amplifier of cognitive abilities and experiential learning. It allows for complex scenario simulations, promotes self-directed learning, and integrates gamified elements that increase engagement for learners and professionals alike.

To accelerate adoption at the industrial level, it is crucial to pair technology with structured training programs and supportive communities. The combination of accessible tools, awareness, and digital literacy creates an ecosystem where XR can truly scale and become a standard in training pathways. I also believe it is essential to develop environments and programs that are coherently integrated into political frameworks and education policies, ensuring sustainability and scalability of XR initiatives but also long-term systemic support.

[mbesenzoni]

As someone who actively develops these applications and helps SMEs deploy them, I am a huge fan of this policy recommendation. Moving public funding away from closed proprietary applications to open no-code tools targets the exact production bottleneck we see in the field every day. The massive cost and time involved in custom Unity or Unreal development is usually what kills SME projects before they even start.

However, removing the coding barrier is only the first step. Once an educator logs into an intuitive platform, they immediately face the asset pipeline problem. They still need high-quality 3D models and environments to actually populate their scenarios. This is where tying these mandated tools into resources like the XR2Learn marketplace becomes vital. Pointing SMEs toward ready-to-use assets, especially the free ones available there, gives them the actual building blocks they need to hit the ground running without having to hire a 3D artist.

For this policy to be a long-term win, these funded tools also need to guarantee hardware longevity. Baking in open standards like WebXR or OpenXR ensures SMEs will not have to rebuild their modules from scratch every time a new headset is released. Furthermore, as a developer, I strongly suggest these platforms include developer escape hatches. Allowing technical teams to easily inject custom scripts when a company outgrows the out-of-the-box features will prevent software dead ends. Funding the creation infrastructure instead of just the final content is the smartest way to scale the whole ecosystem.

[xr2learn-admin]

Policy Recommendation #2

The proposal is to integrate affective AI standards into the EU AI Act’s education provisions:

• European and national authorities should develop technical standards for emotion-aware AI in learning, covering transparency, data minimisation, and learner consent.
• The goal is to enable responsible use, not to ban these tools, so proven AI systems can improve learning while respecting privacy and ethics.

---

Evidence from XR2Learn

• XR2Learn developed open-source AI models that infer learner states (engagement, anxiety, boredom) from voice, physiology, and body tracking.
• Models use self-supervised learning, needing minimal annotated data, making them practical for real adoption.
• Personalisation is grounded in Csikszentmihalyi’s Flow Theory, dynamically adjusting challenges to match skill and emotional state.
• Validated in VR environments, showing better engagement and learning outcomes.

Without proper standards, these benefits could be blocked. A standards-based approach allows responsible, scalable deployment, fully aligned with the EU’s human-centric AI agenda.

---

Discussion questions:

• Could emotion-aware AI improve learning in your context?
• How can we balance innovation, ethics, and privacy in education?
• Would clear standards make you more confident in adopting affective AI tools?

[GiorgosSap]

This is a really strong take and I like that the focus is on enabling rather than restricting. That’s usually what makes the difference between something staying in research vs actually being used.

As an ML engineer, to me the XR2Learn approach makes a lot of sense. Using self-supervised learning for affect signals is pretty much the only scalable way forward, especially given how hard and subjective annotation is in this space. Also, grounding personalisation in something like the Theory of Flow is nice to see, it’s often missing from purely data-driven systems.

I do think one thing that becomes really important (and maybe worth making more explicit in the standards) is how we handle uncertainty and variability across users. Emotion inference can drift a lot depending on context, culture, even the same user on a different day. So beyond transparency and consent, having some standard way to expose confidence or fallback behaviours would make these systems much more trustworthy in practice.

Overall though, this feels like the right direction, if done well, standards could actually accelerate adoption instead of slowing it down.

[mari.conte]

Really strong proposal, focusing on standards rather than restrictions makes a lot of sense.
The XR2Learn results are compelling, especially with personalisation grounded in Flow Theory. Emotion-aware AI could clearly improve engagement, particularly in immersive or remote learning contexts.
Clear standards on consent, transparency, and data minimisation would definitely make adoption easier and more responsible!

[xr2learn-admin]

Policy Recommendation #3

The proposal is to fund a cross-border XR educational content commons, managed under an open IPR framework:

• The European Commission would establish a publicly funded platform allowing both open and commercial reuse of XR assets.
• Content could include 3D models, simulation modules, software, and full training scenarios.
• Licensing and intellectual property management would be transparent, enabling safe sharing while protecting creators.

Evidence from XR2Learn

• XR2Learn built a marketplace using NFT-based IPR management, letting innovators trade 3D models, software modules, and full XR applications.
• 67% of contributed assets in funded projects were 3D models, showing high potential for reuse.
• A European XR content commons, inspired by Open Educational Resources (OER), could dramatically reduce costs, accelerate adoption, and encourage innovation across borders.

---

Discussion questions:

• Would your team use shared XR assets if they were openly available?
• What would make you confident about licensing, reuse, and quality?
• Could this kind of platform change the speed of XR adoption in education and training?

[mari.conte]

An open XR content commons could lower costs and unlock much faster adoption across Europe.

For teams to engage, clarity on licensing, quality standards, and content traceability will be key. The XR2Learn approach to IPR is a promising step in building that trust.

[enrique.hortal]

I strongly support this direction. From an affective computing perspective, the question is not whether to use emotion-aware AI, but how to use it responsibly. Emotional states like boredom or frustration clearly affect learning, and systems that adapt in real time can improve both learning outcomes and inclusion. At the same time, we cannot forget the inherent privacy risks. Emotional data is highly sensitive, so strong standards on data minimisation, transparency, and consent are essential to build trust.

I also want to highlight the value for learners with special needs. In some cases, even human teachers may struggle to recognise students’ affective states. Well-designed AI can provide useful additional signals and support more personalised learning.

I agree as well on the importance of handling uncertainty. These systems should be transparent about confidence and avoid strong decisions based on weak signals. Incorrect adaptation can lead to detrimental outcomes, disrupting the learners’ experiences.

Overall, focusing on standards rather than restrictions is the right approach to enable responsible innovation.

[Sofiya]

Very good point, I agree with this one!

[xr2learn-admin]

Policy Recommendation #4

The proposal is to mandate evidence-based, affect-aware instructional design grounded in interoperable EdTech standards for publicly funded XR learning programmes:

• European and national authorities should require projects to adopt validated frameworks like XR2Learn’s open-source educational model.
• Instructional design should dynamically adjust challenges and tutoring based on learner skill and emotional state, ensuring high-quality learning experiences.
• Interoperable standards would allow consistent, scalable, and reusable XR educational content across sectors.

---

Evidence from XR2Learn

• XR2Learn applies Csikszentmihalyi’s Flow Theory and Vygotsky’s Zone of Proximal Development to adapt learning in near real-time.
• In the project’s open calls, projects required to adopt the framework delivered better pedagogical quality, wider learner reach, and validated industrial training applications.
• The framework is production-ready, reaching TRL 7 and MRL 5, showing that standards can improve both educational outcomes and scalability.

---

Discussion questions:

• Have you experienced XR training that felt engaging and adaptive?
• Would requiring validated frameworks improve XR adoption in your context?
• What barriers might arise when introducing pedagogical and technical standards?

[Sofiya]

It is on Europe to catch up with such topics as XR and I thing education can be a flagship here. That is why I see XR2Learn’s vision so important!

[filisia]

It is very common for those of us on the technical side to underestimate the importance of the educational dimension in our applications. However, the XR2Learn Open Calls have provided impressive evidence of how a dedicated educational framework can significantly boost user outcomes, specifically showing over 20% higher performance. Validated cases like this are exactly what is needed to move XR beyond a niche tool and drive widespread adoption of the technology across all sectors.

[g.minerva]

I particularly like the idea of affect-aware adaptation—if done well, that’s where XR can really move beyond traditional e-learning and deliver more personalized, effective experiences. The possibility to adjust content based not only on skill level but also on the learner’s emotional state is especially compelling.

At the same time, I wonder how feasible this is in practice, particularly for smaller organizations with limited resources. Implementing reliable emotion detection, integrating adaptive logic, and maintaining these systems over time could be quite demanding from both a technical and financial perspective. That said, with the support of European projects and national initiatives, this becomes much more tangible and potentially impactful for organizations like ours.

There’s also the question of data privacy and ethics. Collecting and processing affective data introduces additional complexity, especially in the EU context where compliance requirements are strict. Ensuring transparency, user consent, and a clear value for learners will be key.

Overall, the concept is very promising—thanks for sharing!