A Semantic Future for AI

In our modern information society, people need to manage ever-increasing numbers of personal devices and conduct more of their work and activities online, often using heterogeneous services. The amount of information each individual has to process is constantly growing, making this information increasingly difficult to control, channel, share, and constructively use. To mitigate this, computing must become much more human centered - for example, by presenting personalized information to users and by respecting personal preferences when controlling multiple devices or invoking various services. Appropriate representation of the information's semantics and of the functionality of devices and services are critical to such personalized computing. Symbolic artificial intelligence techniques provide the method of choice for the required semantic-representation and reasoning capabilities. The challenge for symbolic AI is to support large-scale, distributed, dynamic knowledge bases enabling highly adaptive and evolving systems. AI must also look to specific application contexts and develop real-world solutions for problems in those domains. We present some examples of such application contexts