Language is the fabric of technical communication, enriched by images, videos and sound recordings. High quality textual or multimedia content may be thoughtfully composed, carefully organized and transparently partitioned, in the world of digital information processing, it is just unstructured data. The content may be generally accessible or effectively adapted to particular users, for the machine there is just one way to make sense of the digital mess and this is AI. But if AI can make any sense of content it should be able to support its production.
In my talk, I will first share some observations of the state of the art and major obstacles for AI technologies. I will also discuss potential and limitations of AI applications in content production and management. Two core questions are: (1) Can and should AI help the content developer and (2) Can and should the content developer help AI.
There is a trend towards semi-structured content, enriched by machine-usable metadata. But how much structure is needed and how much structure is possible without damaging the original purpose of the content or driving up production costs? We will look at different approaches of mixing structured and unstructured information and hopefully trigger a discussion of content as another interface between people and machines.