Maximizing the Power of Generative AI in Corporate Practice with Legal Certainty

Notes on practical application in the corporate context

Despite the risks outlined, generative AI tools can already be used in corporate practice with comparatively high legal certainty if, in particular, the data protection and the Protection of trade secrets get noticed. By collaborating with AI vendors specializing in data security, using GDPR-compliant AI models, and adapting open-source models, companies can maximize the power of generative AI while minimizing risk and complying with regulatory requirements.

Use of non-European service providers

The technologies of non-European AI providers such as OpenAI (the company behind ChatGPT) should only be used in the corporate context if it is ensured that the data entered is both data protection compliant and does not represent any trade secrets. In this regard, for example, using it to generate marketing and information materials is unproblematic. In the context of controlling and finance processes, however, only limited use cases can be found that meet these requirements. In any case, technical protection should be used, e.g. by checking the input data before sending it from your own IT infrastructure, which is relatively easy to implement.

Use of GDPR-compliant AI models

The German company Aleph Alpha has focused on providing explainable and trustworthy AI models that GDPR Compliant are. The use of these models ensures compliance with the GDPR and offers flexibility in the possible uses. In this case, too, you should refrain from entering data that constitute trade secrets. Aleph Alpha offers the possibility of integrating the language model completely independently into the infrastructure of your own company for such applications. This ensures that the data does not leave its own servers.

Use of open source models

To protect trade secrets and prevent data leakage, companies can Open Source Models use, which can be individualized and also installed on the company’s own infrastructure. This approach allows organizations to remain in control of their protected information while benefiting from generative AI capabilities. It is critical to note that the capabilities of these applications, despite immense improvements, are not yet up to the proprietary solutions like ChatGPT or Bard reach. However, the rapid development of the past few months suggests that this will be the case in the medium term.

* Conversational finance means the use of voice computers or chatbots in the finance area to automatically answer common inquiries and analysis requests.

the authors

Prof. Dr. Mike Schulze

Vice President for Research and Practice Transfer and Professor for Controlling, Accounting and Financial Management at the CBS International Business School on the Mainz Campus. In addition, he works as an external lecturer in the field of executive education at the EBS University of Economics and Law and the Frankfurt School of Finance & Management. His current research and consulting focuses include trends and developments in the CFO area as well as digitalization of corporate management.

Felix Broßmann

Director Intelligent Automation at SKAD Schulz and Krill Advisory GmbH in Frankfurt am Main. He advises companies and authorities on the selection and implementation of suitable AI technologies. His current research focuses include the application potential of AI technologies in medium-sized companies and the effects of AI regulations on organizations.

Prof. Dr. Bernd Wallraff

Professor of business administration, in particular business psychology at the CBS International Business School on the Mainz campus. He is also a lecturer in Executive Education at the Frankfurt School of Finance & Management. One focus of his research is in particular the interface between artificial intelligence and leadership. As a certified coach and organizational developer, he advises managers in the areas of innovation culture, agility and digital leadership and accompanies strategy and change processes.