Generative journalism


Generative AI produces something. Without evaluating quality or quantity, GenAI has an output. The prerequisite is that the construct "artificial intelligence" has been trained, i.e. fed with a lot of data in order to form a picture of the "world". Providing data for a chatbot, which in turn is based on machine learning and GenAI, is surprisingly easy: Google has a free AI tool (beta stage) that can record up to 50 sources. I fed the machine with four years of E3 magazine and the link to our website.
After a few minutes of "thinking", the first answers were provided. What the Google AI machine offered was surprisingly concise and almost error-free. Naturally, it was not storytelling, but more of a tabular list. The AI answer was more like a data sheet and less like an SAP community report. But it was useful for a quick overview.
The definition of generative grammar, which describes human language using mathematical logic, algorithms, automata and psychology, is helpful in understanding these processes. Generative grammar theory explains how it is possible for a source to generate an infinite number of sentences in a language by mastering finite sets of rules. Generative AI therefore has the potential to change work. It can expand skills by automating some of the work. In other words, generative AI can take over tasks that currently account for 60 to 70 percent of working time. F
n the past, experts predicted that IT would only be able to automate half the time at most. But the technology is getting better and better and can do more and more. A big part of this is that AI can understand natural language usage. This is important for work that accounts for 25 percent of total working time. So GenAI has a greater impact on knowledge work, which is associated with occupations that have higher wages and educational requirements.
Will I now be unemployed as a journalist? The question is justified and complex. In his Critique of Pure Reason, Immanuel Kant examines the extent to which our experiences really bring us new insights. According to Kant, we always need both experience and reason. According to the empiricists, pure intellectual knowledge appears to be insufficient. So what can generative AI do? What opportunities does journalism have?
GenAI is based on existing knowledge and can deliver surprising results through systematic analysis. This is by no means new! The findings of GenAI may be enlightening for less well-informed people. But Kant would classify them as largely tautological, just as he was of the opinion that mathematics does not really produce new findings, but at best tautologies of already known ones.
Journalism is not just storytelling, but also an intellectual activity. Thus, there is hope that generative journalism will really generate new insights. I will start cooperating with GenAI to reference and verify my own findings. I am looking forward to a new year with many experiences in the SAP community and wish the E3 readers insightful minutes with our magazine.