The Potential and Problems of Generative AI
Thierry Kremser, PwC | Tate Teague, Revantage | Frank Roessig, Proximus | Andreas Braun, PwC
AI is the third digital revolution after the internet and the cloud, said panel chair Thierry Kremser, a partner with PwC. Specifically, generative AI is enabling us to interact in productive new ways with information and automate tasks, and the technology is only in its infancy. What should directors know and do?

A boardroom topic

“AI should be a topic in the boardroom, and don't look at it only from the compliance perspective,” said Andreas Braun, Artificial Intelligence Leader at PwC. “Keep the opportunities in mind for your own organisation, or for organisations that you might see as a potential investment targets, because they tend to be growing faster,” he added. “Boards should raise awareness and support senior management regarding AI’s potential transformative impact, and helping them understand what AI can or cannot do,” said Tate Teigue, associate vice president data science of Revantage. 

As for how this awareness could be raised, Frank Roessig head of digital finance solutions cloud data & AI at Proximus advised seeking hands-on experience. “Get involved in a project,” he said. “Come to see the different milestones and results, the ideation, the production environments, and get a feel for the process.” He also recommended board seminars with specialists to help directors familiarise themselves and ask questions about the topic.

Achievable use cases

As to some of the use cases, the panel cited the example of a client who used generative AI to cut the time to analyse 300 contracts from ten days to one, including facilitating the review and comparison of documents and legal texts. Another was able to consolidate data more quickly from multiple sources for reports. In the real estate sector, AI tools are also helping tenants and owners communicate more effectively. The creation of FAQs, chatbots, fraud detection systems, and anomaly detection were also seen as low hanging fruit. Real-time translation also has clear potential.

“This is about having augmented human beings, not substituting them,” said Mr Roessig. Specifically, this could feature tools such as intelligent document processing, the Microsoft Copilot chatbot tool, and proprietary chatbots that have been trained on the organisation’s knowledge base. Generative AI is now widely used by software developers, with a large majority of research proposals also being supported by these tools. 

Staffing and regulation

On the talent required, “you need subject matter experts who can design systems and assess if the AI is actually working,” said Mr Roessig. “So the talent pool is broader than people think,” he added. Also, many tasks can be made more efficient with minimal AI support, so in these cases there is not necessarily a need to hire a development team. Future innovations such as Microsoft embedding generative AI into their Office suite will also drive things forward.

The panel also explained that boards also need to get up to speed with the coming regulatory environment, particularly the EU’s planned AI Act. This will apply to developers, importers, distributors, and users of AI in the European market, with oversight dependent on the level of risk implicit in the application being used. Consideration also needs to be given to how AI-augmented processes and products can be explained to auditors and regulators.

Mr Braun said the companies he works with in Luxembourg have shown relatively mature levels of data governance and data management practice, with businesses willing to purchase AI tools and experiment with them. Mr Roessig noted that the technology is moving fast. “Keep an eye on it because what we have today is just the start and has nothing to do with what we will have in the future.”

ESG Oversight - The Key Role of Board members
Geoffroy Marcassoli, PwC | Kenny Panjanaden, PwC | Jane Wilkinson, Non-Executive Director | Claire Collet, UBP Asset Management (Europe)