Schalast | Artificial Intelligence
Advances in artificial intelligence (AI) research require innovative adaptability in the legal field as well. On April 21, 2021, the EU Commission presented the draft of a European AI regulation - this is the world's first legal framework for AI from a geopolitically relevant standard-setter. The regulation is intended to promote AI development on the one hand, while ensuring a high level of protection for public interests and creating trust in AI systems on the other. The regulation includes bans on certain AI system practices, the definition of requirements and obligations for AI systems with high risk, as well as transparency obligations for certain other AI systems.
Application examples are diverse: For example, when it comes to AI in connection with autonomous flying machines, vehicles or RoboCops, classic legal concepts such as strict liability in the event of dangers due to new technical developments are no longer satisfactory today. AI requires a more comprehensive economic law consideration. Should control systems "sacrifice" the driver if an accident is imminent or should they avoid an approaching automobile and kill children playing on the side of the road? Which “programming,” because AI is ultimately only software that has been improved primarily with Big Data, is preferable here? What is the situation in terms of AI tort and product liability? This becomes all the more important as the expectation is confirmed that AI will be incorporated in more and more products used by everyone in everyday life. Particular problems revolve around special questions: For example, can a manufacturer exculpate himself by stating that the AI had developed itself further and it was not foreseeable it would cause damage? It is clear possible uses for software solutions that are able to recognize patterns on the basis of a large number of data and make decisions on this basis are relevant for fields as diverse as
- editing contracts for regulatory change projects,
- analyzing court judgments,
- calculating the success chances of litigation strategies or
- creating psychological profiles in personnel interviews.
In sanction law, smart sentencing models are causing a stir. The financial supervisory authority has also consistently recognized AI's potential and is actively shaping the debate.
Inasmuch as AI’s current state of the art regularly requires the evaluation of gigantic databases, special attention is paid to handling personal data. Further restrictions for the AI industry are to be expected here, at the latest as soon as data protection authorities have fully grasped the issue in technical terms and its huge extent.