Undeveloped and incomplete data markets

Challenge: Undeveloped and incomplete data markets

• The rules for data utilization are inadequate for the production and use of learning services. The lack and ambiguity of standards and rules lead to data silos of different parties. For example, the Ministry of Education and Culture and the Finnish National Agency for Education have been wary of changing existing rules or creating new common rules, such as requirements for common information models and interfaces. To develop platform-type teaching solutions, rules that would increase the potential for using data would be needed, such as rules on 1) what data could potentially be used for, 2) what kind of teaching data could be shared and what would be are privately owned, 3) who is the owner of the data, 4) who has the right to use the data, and 5) who supervises the sharing of the data. For example: if a user profile is created for a student, who owns the data, with whom may it be shared, and can anonymised data be shared freely or is a permission required? 

  Additionally, there is no clear division between data concerning students, learning, teaching and learning materials. The use of student data is governed by privacy and data protection regulation, for example. Similarly to health data, learning data is also sensitive. From the perspective of the public sector, sharing them involves risks for the data provider (protection of the individual). On the other hand, learning data can provide indications of competence levels, suitable learning methods and objects of interest. Teaching data are managed by both the private and the public sector. The private sector sees teaching data as a factor of its competitiveness. The data are closely linked to the learning solution offered by an actor, which means that actors are not willing to share them. While similar restrictions do not apply to learning materials, no effective business models or services have emerged for the reuse of learning materials. For example, teaching data related to a specific course may include behavioural data as well as information on solutions that do or do not work, and optimisation of the level of difficulty. Similarly, data concerning teachers can be used to promote teacher development, and teaching portfolio data can be used for the planning and management of teaching. The quality of data and its regulation is also an important factor. For example, teaching content must be linked to the curriculum, be up to date and be objective. The rules for combining different data are also inadequate. For example, can learning data be combined with other behavioural data? Data protection and information security practices need to be clarified.

• Unwillingness to share data and be open. The data in the education sector have become siloed and do not move between different services. As regards the public sector, data sharing is hindered by concerns about information security and data protection, the fear of private actors, avoiding dependency on suppliers, and regulation/legislation. In addition, no centralised data repository exists in this sector resembling the Kanta repository of the health care services. While legislation has rightly been passed on the use of data, rather than owning an individual’s data, the authorities only manage it reliably. Publishers, in particular, are at an advantage in managing learning material data because of their copyrights, even if the materials were acquired using public funds. Other business actors, including Google and Microsoft, have a competitive advantage based on retaining part of the data for themselves. Innovators do not have the same access to student, teacher and material data as established actors.

Recommendation: Performance of the data market in the education sector should be improved through common rules

• A common data roadmap for the field of education and rules for the use of data should be created. At national level, a joint plan for sharing and using data, including standards and good practises, should be created for public and private actors in Finland. The data roadmap will clarify the development stages of defining the data and information sharing architecture, including rules for sharing data concerning students, learning material and teaching, and the roles related to maintaining the data and information architecture. The rules must take a stand on who owns the data, who has access to them and who monitors the sharing of data. The data needs of start-ups should also be taken into account when creating rules. Platform-type solutions, including Udemy, Coursera, opetus.tv, aoe.fi and Freed, make it possible for anyone to produce and distribute learning materials for the use of others. The roadmap should support the finding of an overall solution for developing the data market in the education sector, rather than only different parts of the chain (data concerning students, learning, teaching and learning materials). This development could be supported by educational data brokers, whose role should be clarified: for example, the practices for sharing data in the Koski data reserve or the Wilma system with those creating complementary systems. Finding overall solutions can also be promoted by modifying the conditions of funding systems (e.g. joint applications by several municipalities).
• Responsibility for managing learning data should be assigned to the individual, and a minimum level of learning data to be shared should be determined. Learning data describes the fundamental characteristics of a person and their learning path. Everyone should be allowed to decide how their learning data are managed and shared. However, in order to clarify the rules of sharing data, a framework based on common rules should be defined. It is best to avoid ecosystem structures in which all learning data must be shared by default when using the service, or in which the learning data are always linked to an individual/identity. This framework should be jointly built by producers of learning solutions (both large companies and start-ups), municipalities and the Finnish National Agency for Education. If the data were generated in a directly usable format, a more comprehensive view could be obtained of the student and their entire learning path (how they learn best, their use of digital content). The data could, for example, be used to develop curricula and better algorithms related to learning technology, and to identify early indicators for learning difficulties. Google already possesses a more comprehensive picture of Finnish adolescents than Finnish learning solution providers. To create smart learning solutions, innovators need enough users and data and the possibility of combining data from different sources. Experience gained from MyData pilots can be used to manage learning data and pedagogical data. By combining learning data with data concerning hobbies, mobility and health, increasingly targeted and useful services could be offered. When developing smart learning solutions, the individual’s responsibility for managing learning data should be addressed. Data ownership, data analytics and the use of artificial intelligence in learning involve continuous setting of ethical boundary conditions – particularly when the data concerns minors.