Prof. Dr Agnė Paulauskaitė-Tarasevičienė: It’s not enough to have AI – you need to understand how it works

The more we discuss artificial intelligence, the clearer it becomes that it’s not just about what AI can do, but how it works, what principles guide it, and who truly benefits. Today, technological progress is inseparable from responsibility – because AI can help solve global problems, but it can also exacerbate them. That’s why we increasingly ask not “Do we have AI?” but “Is it being developed responsibly – sustainably, ethically, and with a clear focus on human well-being?”

According to Dr Agnė Paulauskaitė-Tarasevičienė, Director of the SustAInLivWork Centre of Excellence and professor at Kaunas University of Technology, one of the most common mistakes is thinking you can simply ‘buy’ an AI solution and it will work on its own.

Sustainable AI is a sign of a society’s maturity

According to Paulauskaitė-Tarasevičienė, explainability is particularly important in areas where AI directly affects people’s lives, health, financial or legal situation – for example, in healthcare, transport safety, credit scoring or recruitment.

“I’ve personally experienced an AI decision that made no sense at all. It was only by finding out how that company’s algorithm worked that it became clear the problem was caused by a very limited feature set – the model simply couldn’t handle unusual situations,” she says.

When an expert understands the principles behind an AI model’s decision, they can judge whether it’s valid, ethical and appropriate for the context. This builds trust in the system and keeps humans at the centre of decision-making.

“XAI puts people back into the decision chain and allows AI to be used as an assistant rather than an autonomous tool – especially in sensitive or high-risk situations. I hope that the ‘glass box’ approach will gradually become the norm, replacing the so-called ‘black box’ systems, and that users will come to expect as a basic right the explanation for a decision, whether that’s presented in text, graphically or digitally,” she adds.

As the SustAInLivWork Director says, AI is not magic – it’s maths, statistics and programming. Explainability is a step towards a more mature, responsible and trustworthy use of AI.

The essence of AI is its impact, not the technology

SustAInLivWork is the first competence centre of its kind in Lithuania, systematically bringing together AI knowledge and skills from four leading Lithuanian universities – KTU, Vytautas Magnus University, Lithuanian University of Health Sciences and Vilnius Gediminas Technical University – in partnership with progressive partners: Tampere University (Finland) and Hamburg University of Technology (Germany).

According to her, the Centre’s main goal is to build a responsible, sustainable AI ecosystem that serves people – applying AI not just for technological progress but to improve society and the environment. Its mission goes much further than developing solutions – it’s about systemic transformation: changing mindsets, developing skills and influencing policy.

The solutions created by SustAInLivWork are all about real impact. For the Centre’s researchers, technological progress matters only if it is matched by responsibility, transparency and public trust. Their work is firmly rooted in AI ethics, explainable AI and sustainable development.

“Our principle is simple: not AI for its own sake, but AI for impact. It must be ethical, explainable, transparent, compliant with the EU AI Act and build public trust in technology,” Paulauskaitė-Tarasevičienė says.

Testing before investing

SustAInLivWork’s combined expertise gives businesses not just access to advanced AI technologies but a strategic opportunity to grow responsibly, stay competitive and prepare for the challenges of digital transformation.

“Today, the Centre can help companies improve efficiency, reduce risks, make data-driven decisions and comply with increasingly strict EU regulations. All this is possible through the services we provide,” says the SustAInLivWork Director.

Shortly, companies will be able to use SustAInLivWork’s lab and testbed infrastructure – to test AI solutions in real or simulated conditions (‘test before invest’), verify systems and get independent expert assessments.

“Companies will be able to become early adopters of new AI solutions, try out innovations before they reach the market and gain a competitive edge. They’ll also have access to valuable research-based multimodal datasets, help define challenges, generate new ideas, create start-ups and join national and international research and innovation projects,” Paulauskaitė-Tarasevičienė adds.

SustAInLivWork offers not just services but a co-creation space – a place to grow, experiment and develop AI solutions in collaboration with academia. It’s not just about deploying AI – it’s about understanding it, applying it responsibly and creating long-term value based on trust.

“In the future, this combination will only become more important. AI will no longer be an isolated technology – it will be an integral part of sustainability, work organisation and innovation strategies. It’s a new mindset that helps businesses grow stronger economically while creating greater value for society,” she concludes.