Hamburg clearly an AI hotspot, says Alois Krtil

Hamburg News: To what extent will AI influence the German economy and our lives in general on a scale of 1 to 10?

Alois Krtil: If we think of a five-year period, I would say nine. By then, AI will be playing a role in almost all parts of the economy and society. AI is already being used in many sectors from navigation and translation programmes to voice assistants like Siri or Alexa and predictive maintenance. And businesses in particular will adapt AI applications on a large scale. Not necessarily in the core business, but indirectly, e.g., to optimise procedures, to save costs and resources, and in customer and employee relations and in terms of further training or employee recruitment.

Hamburg News: A majority of German companies see AI as an opportunity, but only 9 per cent actually use the technology, a Bitkom survey in September 2022 found. So is there still room for improvement?

Krtil: ​​​​I expect to see a significant increase in the number of companies using AI. The pace of the development is unstoppable and the number of practical use cases is increasing rapidly. Companies are under growing pressure to act and to avoid falling behind. At ARIC, we notice that in the rising number of enquiries. More and more companies want to see whether and how the use of AI is worthwhile. We can offer more suitable programs thanks to the European Union's EDIH funding scheme to boost the digital competitiveness of SMEs and management in particular. These include information, training events and workshops in which we build prototypes with companies or develop 'Proof of Concept' (PoC), i.e., AI test scenarios for management.

Hamburg News: Start-ups use AI extremely actively. Data analysis and AI are the top two technologies in German start-ups, according to a recent Bitkom survey. What is the situation like among Hamburg's 700 or so start-ups?

Krtil: AI plays an essential role in more than 120 start-ups in Hamburg putting the city in third place after Berlin and Munich. If we add the start-ups using some kind of AI application, the figure comes to three quarters of all start-ups. Thus, AI is the key, most widespread technology among start-ups. Hamburg is clearly an AI hotspot.

Hamburg News: Chat GPT has triggered a media hype. Is this interest noticeable in the economy and local government?

Krtil: Absolutely. We designed e.g., an online workshop on Chat GPT for management staff in Hamburg and  for 20 participants initially. We had over 1,000 registrations in a very short time and put it down to a technical error. But the interest was definitely genuine. The same applied to another course to show school principals how to use Chat GPT in the classroom. One third of all school principals in Hamburg enrolled for the course. And enquiries from business and associations have not stopped. I have never seen anything like it and this reflects the enormous need and great interest. Of course, we will offer follow-ups and the topic will play a role at the AI Summit and at the prompt-a-thon that ThIS! Interface Society is offering during this year's fAIstival.hamburg.

Hamburg News: What is a prompt-a-thon?

Krtil: A prompt-a-thon is a practical AI workshop, based on a hack-a-thon, in which we try out generative AI such as Chat GPT for texts or Midjourney for images and show what it is capable of and where its limits lie. We have noticed keen interest in it and for good reason. AI is one of the biggest catalysts of growth as its impacts the entire economic value chain from development to sales and marketing as well as customer contact.

Hamburg News: AI certainly offers many opportunities. However, ARIC stressed the need for a responsible approach to the technology. What is happening in Hamburg in terms of 'Responsible AI'?

Krtil: Plenty, In June, the City of Hamburg, the auditing and consulting firm PwC Germany and the certification company Dekra founded CertifAI, making it the very first of its kind for AI applications. The new AI certification includes the testing, certification and development of technological solutions for automating these procedures as so-called cyber-physical systems (CPS). This joint venture anticipates the requirements of the upcoming EU AI Regulation and is an essential part of the goal of 'Responsible AI' - incidentally, a field in which Hamburg is a pioneer.

Hamburg News:  Can you give one or two examples?

Krtil: At the end of 2018, experts in various fields met in Hamburg's City Hall for a symposium. The result was "Nine theses on opportunities and risks, democratic legitimacy and constitutional control in the algorithmisation of the administration". The theses, written by Friedrich-Joachim Mehmel, former President of Hamburg's Constitutional Court and the Higher Administrative Court and now LawCom Institute, and Professor Wolfgang Schulz, Director of the Hans Bredow Institute Hamburg, remain valid in terms of establishing guidelines for Responsible AI.

Another example is the partnership with the Fieldfisher law firm as well as other corporate partners and scientific institutions to develop viable, interdisciplinary concepts for the practical implementation of Responsible AI.

Hamburg News: A recent survey by the Bertelsmann Foundation has noted public uncertainty caused by disinformation and fake photos or notorious deepfakes. How problematic is this development?

Krtil: It definitely is becoming increasingly difficult to detect deepfakes. That's a real cat-and-mouse game. As soon as software emerges that exposes deepfakes, a counter-software that creates even better deepfakes surfaces similar to the cybersecurity situation. So we have to keep up our efforts to stay ahead in the race for better tools.

Hamburg News: Lastly, do you have any expert tip on how to recognize deepfakes?

Krtil: Details like hands are a good indicator. They are often blurred in a deepfake or there are too few or too many fingers. Basically, a healthy portion of scepticism. The less plausible the photo, the more likely it is to be a fake. Such photos can frequently be unmasked by searching the internet. Many deepfakes are created simply because someone likes the technology and the makers are proud of them. However, deepfakes about wartime incidents are far more dangerous. We can only hope that the tecnhnology can fend them off. But education on the subject and more data literacy is vital. Data literacy should be taught as early as possible.

Alois, many thanks for the inspiring talk.

Interview by Yvonne Scheller