Advancing responsible use of AI in education and research

Smart use of AI strengthens Finnish education, research and skills development

The AI transformation is fundamentally reshaping society. Its impact is particularly strong in education and research, where it is changing processes, methods and ways of working. At the same time, education and research play a vital role in developing AI, understanding its impact and building new expertise.

AI offers significant opportunities for science, innovation and learning, but it also brings risks and challenges. For this reason, it is important to understand, anticipate and manage the transformation responsibly.

AITO framework supports strategic adoption of AI

The AITO framework (AI in Innovation, Research and Education), developed by CSC, helps higher education institutions, research organizations and other public sector actors map out and understand the implications of AI for their operations and lead AI adoption in a responsible and impactful way. 

Managing this transformation requires a holistic approach that considers ethical, technological, social, environmental and economic perspectives. The AITO framework helps organizations assess both the short-term and long-term impacts of AI adoption, from the implementation of technologies and tools to changes in research, learning and working practices, the creation of innovations and entirely new ways of operating.

Developing a shared understanding of the situation is essential. Using the AITO framework, CSC has supported the leadership of Finnish universities and universities of applied sciences strategically plan and lead the adoption of AI.

AI platform for secure and controlled use of artificial intelligence

CSC and the Norwegian organization Sikt have launched a long-term partnership to promote the safe and responsible use of AI in education, research, culture and public administration. As a first concrete step, CSC and Sikt are developing an AI platform to meet the needs of Finnish public sector organisations.

The AI platform enables organizations to use different large language models and data sources securely and in a controlled manner, without each organization having to solve the same challenges independently. A shared central platform with identity and access management ensures a high level of data protection and information security.

Users can choose from a curated selection of language models and data sources according to their needs. Prompts remain under the user’s control and are not used to train commercial AI models. The AI platform can be integrated into other services, such as learning environments, and it enables the secure development and sharing of users’ own applications and AI agents.

The AI platform is technology-independent and supports European digital sovereignty by enabling the adoption of also local language models.

The AI platform is already widely used within the Norwegian education sector, and pilot projects in Finnish higher education institutions and schools will begin in autumn 2026. These pilots explore a variety of use cases, including strengthening AI literacy, supporting teaching through language models, and improving the efficiency of research and development activities.

Leveraging data, analytics and AI to advance the education sector

CSC supports the Finnish education system by producing reliable and up-to-date information across all levels of education. This information is used to improve educational quality and support funding decisions. We help education sector organizations understand changes in their operating environment, identify emerging trends and phenomena, and develop their services and operations.

We enable our customers to develop data-informed insights and assess how data, analytics and AI can support their goals, use cases and development initiatives. Examples of approaches we use include:

AI pilots

We conduct a range of AI pilots to support the development of existing services and to explore new opportunities for leveraging AI. Through these pilots, we work together with our customers to explore whether a language model or machine learning solution can address their specific needs and perform effectively with the available data.

For example, pilots can be used to assess whether machine learning tools can automatically classify textual material or whether a large language model can generate reliable responses for users from the given data.

Recommendation algorithms

Machine learning models can be used to build recommendation systems. CSC has contributed to the development of a tool for a nationally used service that can suggest relevant content based on user input. In Retrieval-Augmented Generation (RAG) implementations, the reliability and relevance of large language model responses are enhanced by providing carefully selected data sources that the model uses when generating answers

Large-scale text analysis

AI can be used to analyze large volumes of text. For example, responses from open-ended survey questions can be analyzed to identify recurring themes and emerging patterns. These insights can help identify weak signals and anticipate future developments in the education sector. All data is processed securely within CSC’s server environments.