Recommendations

"Emerging technologies, including AI, provide opportunities to advance the UNESCO 2019 OER Recommendation. Beyond the creation of new openly licensed content, potential applications include: facilitating the detection of existing openly licensed content online; developing techniques for effective OER curation; translating OER into multiple languages; and facilitating content indexing through the recommendation of descriptive metadata. Other emerging technologies, such as blockchain-based services and applications, could potentially ensure the provenance, integrity, and lawful use of OER. 

Legal frameworks are then not up to date with technological advancements such as these. In the realm of intellectual property rights (IPR), the key to OER, this can lead to ambiguities on what constitutes legal use as well as issues regarding how exceptions and limitations to copyright law may be used even when a work is not openly licensed. Developing clear guidelines and policies that address these issues can help protect the rights of content creators, ensure proper attribution, and create innovative technology spaces for the public good.

The Global Digital Compact, and UNESCO’s ROAM-X principles, which are Rights-based, Open, Accessible, and promote Multi-stakeholder Participation based on cross-cutting principles, and in particular Gender Equity, can inform policies for OER by providing a comprehensive framework that ensures inclusivity, equity, and collaboration in the development and implementation of Open Educational Resource. 

These frameworks and principles, alongside the UNESCO 2019 Recommendation on OER, provide a robust foundation for developing OER. The stakeholders addressed in this document are those in the UNESCO 2019 Recommendation on OER^[1] . With regard to Generative AI, it is important to underscore that these recommendations apply both to the inputs and outputs of LLMs. Furthermore, the principle of human centered use of technology prevails in all recommendations. The principles of transparency and knowledge sharing  are fundamental for the implementation of these recommendations. The below guidelines aim to provide actions to harness the opportunities posed by emerging technologies, such as AI, for expanding knowledge sharing and creation through the implementation of the 2019 Recommendation on OER. 

^[1] teachers, educators, learners, governmental bodies, parents, educational providers and institutions, education support personnel, teacher trainers, educational policy makers, cultural institutions (such as libraries, archives and museums) and their users, information and communications technology (ICT) infrastructure providers, researchers, research institutions, civil society organizations (including professional and student associations), publishers, the public and private sectors, intergovernmental organizations, copyright holders and authors, media and broadcasting groups and funding bodies.

Conclusion

The 3rd UNESCO World OER Congress reaffirms the transformative potential of OER as a cornerstone for equitable and inclusive access to knowledge in the digital age. By harnessing emerging technologies, including AI, and embracing open solutions aligned with the principles of Digital Public Goods (DPGs), the global community can foster innovation while addressing challenges related to access, intellectual property, and sustainability.

The commitments made through the Dubai Declaration aim to bolster global efforts to bridge the digital divide, empower diverse communities, and contribute to the advancement of the United Nations Sustainable Development Goals (SDGs). Through collaborative action, we can ensure that OER remains a catalyst for lifelong learning and an enabler of a fairer, more knowledge-driven future for all.

Suggested actions for the five areas
Capacity Building

1. Support the continuous professional development of contributors, content developers, and content providers, on Generative AI projects, regarding the importance of adhering to and respecting open licensing terms of both input and output content. This would also include issues related to training on understanding how commercial AI services impact open content ecosystems, allowing them to critically engage with licensing; and reinforce skills to take an active role in developing specialized, responsible, community-driven large language models (LLMs) for OER.
2. Promote continuous, contextualized and inclusive digital literacy capacity building for OER stakeholders at all levels toallow stakeholders to assess, understand, review, and engage in the responsible development and use of AI andemerging technologies for OER.
3. Support the development of linked technologiesto improve attribution and discoverability of OER. Prioritize data protection, interoperability, and the use of thesauri and controlled vocabularies to streamline metadata across platforms. Focus could be on building interoperable, open-source AI frameworks for OER that promote decentralized control and safeguard content integrity. 
4. Support the development of guidance that promotethe implementation and prioritization of digitally signed works for OER, and their re-use in the training of open AI models for openly licensed learning content. 

5. Implement strategies grounded in human rights that are open, accessible, including those in vulnerable situations, multistakeholder, gender inclusive to ensure protection and respect for user generated data, metadata, privacy and attend to ethical practices and respect copyright rules.
   
6. Policy

1. Advocate for Generative AI platforms to recognize and respect the licenses and authorship of openly licensed content. This would entail incorporating open licensing into the Terms of Use of AI applications, specifying that it is only to be used by humans to generate openly licensed content and implementing robust monitored to ensure compliance. 

2. Advocate for the recognition of open licenses in AI model training, ensuring that both the input data and generated content reflect proper attribution to the original creators. This includes embedding licensing information in outputs and promoting the use of compatible open licenses to safeguard the provenance of openly licensed materials used in AI models. 

3. Advocate for the adoption of machine-readable licenses in metadata to ensure proper attribution of both training data and generative AI outputs to clarify the use of OER as AI training data under existing licenses and how AI systems should attribute this usage

4. Encourage and support research into next-generation attribution systems to enable the tracing of the use and re-use of OER, open solutions and OER as digital public goods, integrating in research elements of Rights, Openness, Access to all, Multistakeholder participation, and Cross-cutting issues to develop transparent, accessible, and equitable systems for attributing, retrieving and reusing O
   
   Ensuring inclusive and equitable access to quality OER

1. Support the development of AI-enabled OER that is accessible in low-bandwidth scenarios and designed to enhance the accessibility of vulnerable groups, including those with disabilities by ensuring its correct integration and readability with assistive technologies,that ensures the highest standards of privacy and data protection during the production, use and sharing of OER.

2. Integrate machine-readable licenses as metadata that include digital identifiers for authorship for authenticating authorshipinto quality criteria for the production of OER and consider it as a criterion for inclusion in OER directories. Quality criteria and policies should emphasize the connection of authorship to real-world identity of authors - both to create incentives for publication and to counter misinformation efforts.

3. Support the translation and contextualization of OER into multiple languages, using AI-related technologies when appropriate, with due attention to the quality of the output of translations and its cultural relevance in collaboration with users’ communities

4. Promote open ecosystems that prioritize the development of digital public goods and open solutions. This includes fostering robust public infrastructure and public-private partnerships, while also supporting novel private initiatives for OER using emerging technologies, including AI, that adhere to the principles of digital public goods and high-levels of openness.
   
   Sustainability Models for OER

1. Support approaches to ensure Interoperability, Intellectual Property Rights (IPR) protection as well as sustainable scaffolding of OER development grounded inthe ROAM-X principles of human rights, openness, accessibility, and multi-stakeholder participation and cross-cutting issues of gender.

International Collaboration

2. 1. Establish regional and international networks to support collaboration to empower community driven OER AI development.These networks should also support ahuman centered use of emerging technologies, including AI, for the implementation of the 2019 UNESCO Recommendation on OER.
   2. Establish mechanisms to support engagement with the wider open community and legal experts on open licensing and intellectual property law to develop guidance and capacity training toensure that emerging technologies, such as those integrating generative AI, adhere to evolving legal terms and address the demands and voices of diverse stakeholders. 
   3. Support cooperation including at inter-regional and inter-sectoral levels amongst Research and Development Centres to develop, and advance responsible AIframeworks as well as support the development of guidance and training for users on emerging technologies and novel methods to promote OER, in light of emerging technologies and AI. 
   4. Support collaborative frameworks so that OER repositories, and other sources of open content, develop and implement policies that prioritize the integration of authorship recognition models, and also clearly define how these works may be processed and used, including criteria for the training of AI models.
   5. Facilitating the development of platforms using AI techniques where stakeholders can create and co-create OER adheringto the 2019 UNESCO Recommendation on OER.