"Generative AI will live up to the hype - provided we use it with clear goals and a high degree of responsibility."

Generative AI is one of the most disruptive trends in technology. It enables machines to independently produce creative content such as text, images, videos or music, revolutionising sectors ranging from art to medicine and industry. 

Dr. Sven Schmeier heads the Competence Centre Generative AI at the German Research Centre for Artificial Intelligence (DFKI) and specialises in the development and research of generative technologies with a deep understanding of machine learning, natural language processing and innovative AI applications. Under Dr Schmeier's leadership, the competence centre not only drives technological innovation, but also addresses the ethical, legal and social challenges of generative AI.

We spoke to him about the work of the Centre, the changing focus of AI, and the potential of generative AI.

Dr Schmeier, can you give us an overview of the core tasks and research goals of the Competence Centre for Generative AI at DFKI? What do you focus on in your work?

One of our most important tasks is the systematic consolidation of research and research results at a central location. This not only gives us a better overview of current developments, but also creates a solid foundation for future research projects. Another essential aspect of our work is to promote the exchange of information between the various research departments, of which DFKI now has 26, spread over 7+ locations. The possibility of closer integration between the various specialist disciplines creates valuable synergies that significantly advance our research.
The realisation of practical projects is particularly important to us. We realise these both internally for various DFKI departments and externally with our partners from industry and science. In this way, we ensure that our research results are not only theoretically sound, but also create practical added value.

To what extent is the work of the Generative AI Competence Centre interlinked with the projects of the DFKI Labs in Berlin? Are there any joint research topics or projects that are particularly noteworthy?

The Competence Centre Generative AI at DFKI works closely with the DFKI Labs in Berlin to promote interdisciplinary approaches in generative AI research. This cooperation allows us to pool different skills and develop innovative applications that are of great benefit both scientifically and practically. One joint project that is currently receiving particular attention is the evaluation of generative AI systems used in various applications. Here we are working on developing new methods for evaluating the performance and safety of such systems in order to improve their applicability and trustworthiness.

The main projects of DFKI and the DFKI Lab in Berlin are described in detail on the official DFKI website. Here you will find comprehensive information about our current research projects, application areas and partners. Colleagues from the Generative AI Competence Centre are actively involved in these projects and contribute significantly to their development and implementation. Through this cooperation, we ensure an interdisciplinary exchange and the targeted further development of generative AI in various application areas.

GenAI is currently attracting a lot of attention. At the same time, sceptics who doubt the economic benefits of generative AI are becoming more vocal. You work closely with industry, what impact do you see? Is GenAI living up to the hype? 

As a scientist in Germany, I see generative AI as one of the key technologies that really has the potential to fundamentally change the way we work and do business. The hype around generative AI is of course huge, and a certain amount of scepticism is not unjustified. However, the question is not whether the technology is economically relevant, but how we can use it in a targeted way to create sustainable benefits.

Here in Germany and Europe, we are well positioned to make effective use of generative AI - but this requires targeted education and the will for economic transformation. While the US is driving technological development, Europe's opportunity lies in strategically integrating the technology into everyday working life and using it productively. It is less about the sheer speed of innovation and more about how well we integrate generative AI into our existing systems and processes.

We are beginning to see the economic impact in areas such as automation and improved decision making, particularly where large volumes of data can be efficiently analysed and interpreted. Many companies I come into contact with in my research and consulting work are increasingly recognising that generative AI can help them to be more flexible and agile. But to really live up to the hype, a massive rethink of education and training is needed. Access to these technologies must be widened so that not only large companies, but also SMEs and public administrations can benefit from them.

From a scientific point of view, I say yes, generative AI will live up to the hype - provided we use it with clear goals and a high degree of responsibility.

What challenges and ethical issues do you see in the further development of generative AI, and how does the Centre of Excellence address these aspects in its work?

The challenges and ethical issues associated with generative AI are manifold and require a responsible and strategic approach. First, there is the issue of transparency and accountability. Generative AI models are often complex and operate as a 'black box', meaning that even the developers may not know exactly how a particular result is achieved. It is crucial to design models and algorithms in such a way that their decisions are at least partially traceable - especially when they are used in safety-critical or sensitive areas. Our competence centre is working hard on research into explainable AI to promote this transparency.

Another important issue is ethical responsibility. Generative AI can easily be used to create misinformation, manipulate media or even targeted disinformation. This threat must be taken seriously, which is why we are also researching mechanisms to detect synthetic content. The aim is to develop technologies that not only generate content, but also recognise responsibility, for example through methods for checking sources or veracity.

A third important issue is responsibility for social consequences. Generative AI is changing the world of work and can automate tasks in many areas that were previously performed by humans. This development raises legitimate concerns about job losses and the need for new skills. At the Centre of Excellence, we are focusing on targeted training programmes to equip people with the skills they need to succeed in an AI-driven world of work. We see it as our role to work with businesses, educational institutions and policymakers to create training programmes and thus build 'AI resilience' in society. This is also reflected in the fact that we founded a spin-off of DFKI in June 2023, the AI Transformation Institute in Berlin, with which we offer in-depth training and further education in the field of generative AI.

In short, we are working to ensure that generative AI is used responsibly, safely and transparently. By working with industry partners, but also by supporting start-ups and projects that focus on ethical AI, the Centre of Excellence seeks to promote not only technological innovation, but also social acceptance and ethical standards.

What does this focus on generative AI mean for the AI sector in general? Do you see a danger of other important AI technologies being pushed into the background?

Or in other words: GenAI is eating our lunch. The current hype around Generative AI has undoubtedly led to a strong focus on it in the AI sector, which brings both opportunities and risks. On the positive side, the hype has greatly increased public and business interest in AI as a whole. Generative AI has made AI visible in a way that hardly any other technology has done before. This helps us researchers and developers get more resources and funding, which ultimately benefits the entire AI sector. There will be more investment, more research and more talent attracted to the sector.

At the same time, however, there is a risk that this strong focus on generative AI will mean that other important AI technologies and methods receive too little attention. Traditional AI techniques such as rule-based systems, decision trees or optimisation methods are still highly relevant and often the better choice for specific industrial applications where accuracy, robustness and efficiency are critical. Especially in safety-critical areas such as industry or medicine, many companies rely on these proven methods.

Overall, I see the hype as a double-edged sword: it accelerates development, but it also carries the risk that we lose sight of the big picture. To ensure sustainable development in AI, we need to strike a balance between new and proven technologies.

How are you personally using GenAI? For which use cases would you have liked to see quantum leaps in technology development much earlier?

As someone who works extensively with generative AI, I naturally use this technology on a regular basis, both in my research and in my everyday life. I find generative AI to be a valuable tool for analysing complex text, creating content and exploring new ideas. Text generation and analysis tools are particularly helpful in the scientific field, as they help us to efficiently process large amounts of data and gain new perspectives on our research questions. I also enjoy identifying new areas of application and following the progress that is being made in areas where AI has been largely absent. 

There are some application areas where I would have liked to have seen today's progress much earlier. A good example is the creation and adaptation of learning materials. In academia, and especially in teaching, it would have been a great advantage years ago to have access to flexible and automated tools that could adapt course content for different audiences. This technology now makes it easy to create content at different levels of difficulty or with different focuses, which is hugely valuable for teaching and learning. Another area is supporting the creative process. In the past, I have often wished for a scientific publishing tool that would help me visualise data and create creative scientific presentations. Today's image generation tools do just that, allowing researchers to quickly create visual content that makes complex concepts easier to understand, even for non-experts. This not only saves time, but also improves the communication of scientific findings.

And there are many more areas that come to mind, but they are beyond the scope of this interview.

Where do you see the greatest potential for generative AI in the future? Are there any industries or application areas that you think will particularly benefit from this technology in the coming years?

Generative AI has the potential to transform almost every industry, but I see particularly high growth potential in some areas. One obvious area is the creative sector, i.e. areas such as media, design and entertainment. Here, generative AI can not only help create content such as text, images and video efficiently, but also enable new creative concepts and formats. Creative agencies, publishers and film studios are already experimenting with generative AI, and this trend is set to intensify in the coming years.

Another area is healthcare. Here, generative AI can be used to analyse medical images, create personalised treatment plans and even develop new drugs. Combined with existing diagnostic and therapeutic systems, the technology could help provide more accurate and personalised healthcare services, which could significantly improve patient care. Although research into AI-generated drugs or therapeutic approaches is still in its infancy, it is already showing promising signs.

Manufacturing and logistics are also high-potential areas. Generative models can be used to simulate and optimise production processes. For example, AI can be used to model entire production lines and supply chains to predict bottlenecks and allocate resources more efficiently. These capabilities will be particularly beneficial to companies that rely on complex, globally distributed supply chains. Generative AI will also play an important role in education. The technology will make it possible to create personalised learning content that adapts to users' individual learning levels and preferences. In a globalised and digitised world, it will be crucial to impart knowledge in a targeted and efficient manner. Here, generative AI offers the opportunity to make educational resources more accessible and motivating.

Another promising area for the future is finance. Banks and insurance companies are already using AI to detect fraud and assess risk. Generative AI could also be used to develop more accurate predictive models to forecast complex market developments or create customised financial products. This technology has great potential to transform the financial sector in terms of automation and customer centricity.

In summary, generative AI will find its way into various industries in the coming years, with healthcare, creative, manufacturing, education and finance expected to benefit the most. The challenge and opportunity will be to use the technology in a way that brings not only economic but also social benefits.

Thank you very much for talking to us.