Software and Artificial Intelligence (AI) are no longer neutral technologies. They shape how we produce technology, energy, manage resources, and make decisions in society. At the same time, they consume significant amounts of electricity, water, and computational resources. This creates a paradox: digital technologies are both part of the solution to sustainability challenges—and part of the problem. This tension is closely related to the Jevons paradox: improvements in efficiency, such as more energy-efficient software or AI models, can ultimately lead to increased overall consumption as these technologies become cheaper, more accessible, and more widely used.
These illustrate, why Sustainable Software Engineering (SSE) is emerging as a key field in both research and university education.
Why Sustainable Software Engineering matters
The importance of sustainability in software can be approached from four complementary perspectives:
- Sustainable Software Engineering – how we design and build software systems that are energy-efficient, maintainable, and environmentally responsible.
- Software for Sustainability – how software systems are applied to address sustainability challenges in areas such as energy, mobility, and climate monitoring.
- Sustainable AI – how artificial intelligence itself can be made more efficient, transparent, and ethically responsible.
- AI for Sustainability – how AI can be applied to support sustainable development, for example in optimizing energy systems or monitoring environmental change.
- Green Software Engineering – How software systems are designed and developed to minimize energy consumption and reduce environmental impact throughout their operation.
- Sustainable Software Lifecycle – how sustainability principles are integrated across the entire software lifecycle, from design and development to deployment, maintenance, and end-of-life management.
These perspectives form a simple yet powerful framework:
we must both make technology sustainable and use it to support sustainability.
This dual role is increasingly visible. Even a single AI-generated output consumes energy and resources, while large-scale AI systems rely on energy-intensive data centers. At the same time, these same technologies can enable solutions for climate change, circular economy, and resource efficiency.
Building a research direction: Sustainable AI
To address these challenges, we have initiated a research effort focused on Sustainable AI. The goal is to understand how AI and sustainability are interconnected across environmental, technical, social, and ethical dimensions.
The research is structured around key themes:
- Environmental and technical sustainability (energy use and carbon footprint)
- Ethical and social aspects (fairness, inclusion, and global impacts)
- Governance and policy (regulation and transparency)
- AI applications for sustainability
- Education and research practices
- Measurement and long-term impact
A central Research Question (RQ) is:
RQ1: How are artificial intelligence and sustainability related, and
RQ: How can they evolve together toward a more sustainable future?
From research to real-world practice
One concrete outcome of this work is an AI-based mobile application designed to support informal e-waste workers and its environment. The application leverages image recognition to identify electronic waste components and provide appropriate safety instructions.
This example illustrates the concept of AI for sustainability:
- It enchances occupational safety in hazardous environments
- It is designed to operate in low-resource settings and on low-end devices
- It demonstrates how AI can create direct societal impact
At the same time, it highlights challenges such as system performance, usability, and ethical considerations—reminding us that sustainable solutions must balance technical feasibility with real-world constraints.
Integrating research into university teaching
A key goal has been to integrate this research directly into education. The Sustainable Software Engineering course at Tampere University was designed as an experimental platform where teaching and research are closely aligned.
The course combines:
- Sustainability concepts in software and AI
- Scientific writing and research skills
Students work on research or programming projects, often exploring topics related to sustainable AI. The course effectively acts as a knowledge production environment, in which students contribute to ongoing research while building and developing their own expertise.
The latest course implementation is now coming to end (April 2026), resulting in 25 research essay submissions. These projects cover a wide range of topics, including energy-efficient software, AI applications for sustainability, and ethical aspects and considerations in digital systems.
This approach has several benefits that highlight skills development, and real-world relevance:
- Students engage with real research problems
- New ideas and prototypes emerge from course work
- Some projects continue into theses or publications
- Students develop practical and research-oriented skills
- Students gain experience in applying sustainability principles in practice
- Learning is closely connected to real-world challenges
In this way, education becomes not only a way to transfer knowledge, but also a way to create it as students actively contribute to research and develop new ideas through their work.
Challenges and future directions
Despite promising results, several challenges remain:
- The environmental footprint of AI continues to grow
- Measuring sustainability in software systems is still difficult
- Ethical and societal impacts require continuous attention
- Bridging research, education, and industry practice takes time
- Lack of standardized metrics and tools for sustainability assessment
- Scaling sustainable solutions from prototypes to real-world systems remains challenging
However, these challenges also highlight the importance of the field. Sustainable Software Engineering is not a niche topic—it is becoming a fundamental requirement for responsible technology development.
Looking forward
The integration of sustainability into software engineering and AI is still in its early stages. There is a clear need for:
- Multidisciplinary research
- Stronger collaboration between academia and industry
- New educational models that combine theory and practice
- Standardized methods and tools for measuring sustainability impact
- Scalable solutions that can move from research to real-world deployment
By connecting research, teaching, and real-world applications, we can move toward a future where software and AI systems are not only powerful—but also environmentally responsible, socially just, and economically sustainable.
Sustainable Software Engineering is not just about improving code.
It is about shaping the role of technology in a sustainable world.
Let’s collaborate
Interested in working together on sustainable software or AI? I welcome opportunities for collaboration in research, education, and real-world applications. Feel free to reach out to discuss ideas, projects, or potential partnerships.
Further information and collaboration requests:
Mika Saari, D.Sc. (Tech.)
Vice Head of GPT-Lab Tampere
Software Engineering, Tampere University
Tel: +358 40 826 2928
About the author
