Trust in artificial intelligence paying off, AI summit concludes

5 September 2024

AI expert Alois Krtil presents three theses on potential of AI - no end to boom in sight

Has artificial intelligence (AI) already peaked and is the bubble about to burst? "No," said Alois Krtil, founder and CEO of the Artificial Intelligence Centre Hamburg (ARIC) during the recently-ended AI Summit held by the Chamber of Commerce, ARIC and the Mittelstand-Digital Zentrum Hamburg. The technology has been hyped since Chat GPT arrived on the market in November 2022. Nowadays, generative AI delivers text and images at the touch of a button, enabling applications unimaginable during the first AI Summit in 2019. However, when Nvidia’s shares, the largest manufacturer of chips for AI applications, plunged in July, some analysts predicted an end to the AI boom. "I did the maths. Nvidia's shares had risen by around 2,000 per cent since the ChatGPT rally," says Krtil. Yet, analysts on Wall Street are still wondering whether huge investments in AI are worthwhile? "The fact that not all investments have been amortised so quickly is quite normal when implementing technological innovations," Krtil stressed, adding: "The bubble won't burst."

Strengthening trust in AI

View of packed AI Summit in Chamber of Commerce, Hamburg — AI Summit in Chamber of Commerce, Hamburg

Yet, stakeholders, technology users and the population remain sceptical about AI. Consequently, building trust in AI was one of the main themes of this year's summit. Mirko Gontek, Senior Consultant at Netlight, a consulting firm focusing on digitalisation, called for the demystification of AI, which is fairly easy: "Many AI projects are not moonshots, but lead to simple but lucrative improvements in companies." And an increasing number of implemented projects are leading to greater confidence based on experience. Yet, this applies only to successful AI projects. Testing and smart regulation are two key pillars to building trust, says Robert Kilian, CEO of Certifai, who came out in favour of risk-based, innovation-friendly regulation: "It's not about blanket bans, but about gradual adaptation to the risks posed by the technology." “Real world AI laboratories provide a controlled and practical space for testing innovations and keeping potential risks to a minimum."

Launch of HHAI-Score

Professor Tilo Böhmann, University of Hamburg, presents HHAI-Score

Building trust through understanding is the approach taken by the new HHAI-Score initiative, which Professor Tilo Böhmann, Vice President Research at the University of Hamburg and board member of The Interface Society (ThIS), presented to delegates Monday (September 2, 2024). "We aim to provide 100,000 residents of Hamburg with the skills needed to deal with AI in future." This will be done in four stages: Learn, try, apply and reflecting. "AI will become everyone’s daily companion," he added. This applies to work in the economy and to knowledge. Kathrin Haug, Vice President of the Chamber of Commerce, cited "The Digital Factor" study done by IW Consult, a subsidiary of the German Economic Institute, on Google’s behalf. "Generative AI could generate EUR 330 billion in gross value added across Germany in future," she noted. Thus, there are billions of reasons to position oneself competitively for the future. Krtil then put forward three theses on how this can be achieved.

Thesis 1: Industrial suitability of AI

"After a highly exciting phase, we have now reached the realisation stage. Now’s the time to synchronise the various AI applications with existing procedures." Given the numerous operational AI tools available, businesses and industry no longer have an excuse not to use AI. However, the summit also showed that this will take time. Asked about the level of AI maturity in the German economy, Professor Sebastian Gerling, Chief Digital Officer at the University of Hamburg, said: "We have noted different stages. AI has arrived in many large companies, but most potential users are still trying to understand and build data."

Thesis 2: Responsible AI to stay

Responsible use of AI is key to trust in the technology and Responsible AI is definitely needed, Krtil stressed. "This issue will remain on the agenda." Europe is pursuing this goal with its AI Act, which took effect on August 1. Regulations on safety and ethics in an AI environment are also being developed in Asia and the United States. "More and more companies are developing their own codes of conduct, to be taken into account from the outset," says Krtil.

Thesis 3: Specialisation and decentralisation

Will all this AI hype lead to another 'killer application'? "It already has," Krtil said, citing the AI tool Alpha Fold, which can calculate 3D models of protein structures. The latest version can predict the interactions of proteins with other biomolecules in human cells with great accuracy. "This is a revolution for developing medicines and the natural sciences," he added. In hardware development, on the other hand, an important step has been taken to meet the growing demand for graphics processing units (GPUs) for AI applications, which the manufacturer Nvidia can barely meet: "AI on chips" is the magic word that could lead to decentralisation on the manufacturer's side. More and more companies and start-ups are launching AI chips. "It no longer belongs to the top dog Nvidia alone. And that means an end to 'the winner takes it all mentality'," Kritl stressed.
ys/sb/pb