AI Legal Day by Digital Beat – Legal Aspects of Artificial Intelligence

Artificial intelligence has become an indispensable part of our (office) lives. It automates processes, analyzes data and supports decision-making in many areas. Similar to the PC, it will become an integral part of our daily work. The growing importance of AI is also reflected in the 2024 Nobel Prizes, which recognize achievements that would not have been possible without AI-based research and research on AI.

The AI Legal Day hosted by Digital Beat focused on key legal issues. Experts from various fields highlighted the challenges and regulations that are relevant for companies and developers. Topics such as copyright, data protection, regulation and the use of chatbots were discussed in detail.

AI and copyright

Dr. Dominik Herzog, a specialist lawyer in media law, among other things, explained the copyright consequences of using AI to create content. The crucial question is to what extent an existing work is used as a basis and whether it has been significantly modified.

If an AI creates an image or text based on an existing work without significantly modifying it, the original copyright remains in place. In such a case, the consent of the copyright holder is required to use or publish the new work.

A specific example was shown using a work of art created by a female artist. This version was compared with a variant generated by DALL·E. Since the AI image showed significant changes and was not directly recognizable as a derivative of the original, copyright does not apply in this case. Such distinctions are crucial as AI-powered creativity is increasingly being used for commercial purposes.

AI and data protection

Christian Solmecke was also a guest and explained the data protection requirements for the use of AI and analyzed the 15-page orientation guide of the Data Protection Conference (DSK).

Companies must clarify a number of questions when using AI:

• What will AI be used for? AI must not be used for purposes other than those originally intended.
• Is the training data protection compliant? If personal data is used to train an AI, all legal requirements must be met.
• How will customer data be processed? Consent is required when personal data is used or stored.
• How are AI decisions made? Transparency is the keyword here. It is important to ensure that it is possible to trace how the AI arrives at its results.

A particularly critical point is the question of whether an AI should be allowed to make decisions about people. In sensitive areas, such as application procedures or granting loans, it is imperative that a human being makes the final decision.

Another point of discussion is the security of open systems compared to cloud solutions. Companies must be aware of the infrastructure they use for AI applications. In addition, it must be ensured that deletion rights are maintained. Removing personal data from large language models (LLM) poses a significant technical challenge. One possible solution could be to use filters to prevent unwanted data use.

Data protection officers should be involved in the development and use of AI processes at an early stage.

Data intelligence and GDPR-compliant analysis

Thomas Klußmann presented practical use cases for the data protection-compliant use of AI in data analysis. One example was the use of spreadsheets with an AI connection to automatically evaluate customer surveys and feedback.

It was shown how ChatGPT can be integrated into existing spreadsheets to efficiently analyze large amounts of data. An extension like GPT for Sheets or integration via a script makes it possible to structure survey results and feedback. The crucial point here is that no personal data such as names or addresses are included. The analysis must be limited to anonymized data only.

EU AI Act – Regulation of AI systems

Moritz Gielen, a specialist lawyer in data protection law, presented the EU AI Act, which defines clear guidelines for the use of artificial intelligence. Companies that develop or use AI-based systems must adhere to certain principles.

• The data basis must be objective and diverse to avoid bias and discrimination.
• AI must not make automated decisions about people, particularly in areas such as human resources or financial services.
• Many companies are developing their own AI guidelines to ensure that ethical and legal standards are met. Large technology companies such as Google and Microsoft already have internal frameworks in place that regulate the use of AI.

The AI Act will have a significant influence on the future development of AI applications. Companies need to address the new regulations at an early stage to minimize legal risks.

Chatbots – Increasing efficiency through AI

Another topic was the use of chatbots, which are increasingly being used for customer communication. They can interact with users both in writing and by phone.

botpress.com has summarized important key figures about chatbots:

• 62% of consumers would rather use a chatbot than wait for a human customer service representative.
• 69% of customers would use a chatbot if they knew it could solve their problem faster.
• Chatbots are the fastest growing communication channel for brands, growing 92% from 2019 to 2020.

It is important that the user always recognizes that they are communicating with an AI. Transparency is an essential factor here to ensure trust. Companies use chatbots specifically to automate processes and make customer service more efficient.

Significance of the legal framework for AI

The AI Legal Day showed that the legal framework for the use of artificial intelligence is constantly being developed. Companies must deal intensively with copyright, data protection, regulation and the use of AI-supported systems in order to meet legal requirements. Various consultants are available to help avoid legal risks and use AI responsibly, whether it be a specialist lawyer, a training company or a data protection officer.