How sustainable is AI? Expert explains opportunities and limits

Digitization for digitization's sake is not enough, you say. How do companies find the right course?

The most sensible starting point is the specific problem in the company. Only when it is clear what goal is to be achieved can it be assessed whether and to what extent digital solutions - for example through the use of AI - bring real benefits. Instead of being driven by trends, companies should formulate their own guidelines for digitalization and sustainability goals. These guidelines also help when conflicts of objectives arise, for example between energy efficiency and cost pressure. Companies would do well to focus on particularly well thought-out projects rather than as many as possible. It is quality that counts, not quantity.

AI systems require enormous computing power and therefore a lot of energy. How can progress and sustainability be combined?

The decisive factor is whether a particular model is at all suitable for the problem at hand. In many cases, smaller, specialized systems perform better than large, expensively trained generalists. The operation of the systems must also be scrutinized: How climate-friendly is the data center used? Is the waste heat generated being used sensibly? Which power source is used? Ultimately, it should always be checked whether the expected benefits justify the resources required. Some applications consume a lot of energy without delivering any clear added value. Such cases must be recognized and avoided.

What does an AI system have to bring to the table to be considered a good example of sustainable technology?

A truly sustainable AI system combines ecological, social and economic responsibility. It should work efficiently, i.e. require as little energy as possible and at the same time generate tangible benefits. Transparency is important here: the data sources and model versions used should be documented and traceable. It also needs clear rules for operation. These include regular checks for distortions, assessments of potential risks and ongoing monitoring of the system. Only when these aspects work together can we speak of a good example of sustainable technology.

In your opinion, what distinguishes a sustainable AI project from short-term digital hype?

Sustainable projects are characterized by the fact that they are based on clear assumptions, are developed in a structured manner and are part of a larger change process. They are not isolated, but are embedded in a meaningful way. Digital hype often lacks a clear objective. Oversized systems are set up that are neither efficient nor useful. Sustainability is also reflected in the integration of employees. If skills are considered, strengthened and integrated, this is a strong sign of sustainable digital projects.

How will the role of business informatics change in the future?

It is important to me that students consider the impact of the technologies they develop right from the start. Every decision - whether for a cloud offering, a storage location or an AI model - has ecological consequences. This is where data expertise comes into play: clean, structured data management is the basis for sustainable AI systems. At the same time, interdisciplinary thinking is becoming increasingly relevant. Developers should not only be technically adept, but also familiar with sustainability, ethics and business aspects. This also includes a solid understanding of regulatory frameworks: What does the EU AI Act require? What obligations does a sustainability report entail? And how can these requirements be implemented in practice?

What should young people who want to enter digital professions know?

AI always has an impact on everyday life: it changes processes, influences resource consumption and can shape entire industries. Whoever develops it bears responsibility. It is therefore important to develop an awareness of social, ecological and ethical issues in addition to technical training. The question of how a system can become smaller, fairer and more efficient is often more important than the question of the latest function. Curiosity and the ability to reflect are just as important as a realistic view of the consequences of one's own work. Not every innovation means progress. Sometimes doing without is the better decision. Those who work on digital systems today are helping to shape the future and can consciously influence it in a positive way.

Prof. Dr. Vanessa Just teaches business informatics at the FOM University of Applied Sciences(www.fom.de) and is the founder and managing director of juS.TECH.