General AI Safety falls under the category of Artificial Intelligence and Digital Society. The term describes measures and concepts intended to ensure that particularly advanced, general artificial intelligence (AI), so-called General AI, poses no risks to people or society.
While today's AI is mostly specialised for specific tasks, General AI aims to solve tasks just as well, or even better than humans, across many different domains. For such systems to act reliably and safely, researchers are developing methods to better control their goals and decisions.
A simple example: Imagine a robot assistant helping with care in a hospital. Without general AI safety, it could misunderstand instructions or make dangerous decisions, such as dispensing the wrong medication. With appropriate safety measures in place, the AI would ensure that human interests and rules are always followed.
General AI safety is therefore an important topic to ensure that future, increasingly intelligent AI systems benefit humanity and do not pose any dangers. The goal is to bring technology and society together responsibly.
Deutsch 
English (UK)