Different ways to use AI in technical communication

Thank you to everyone who joined our AI event and a big thank you to our speakers and our hosts from GE HealthCare. We had an interesting afternoon with different presentations focusing on the use of AI in technical communication.

AI and the Future of Subtitling, Peter Bjerre Rosa, DR Media Production Text

Peter told about how artificial intelligence is reshaping both live captioning and prepared subtitling, with real-world examples from Danmarks Radio (DR).

He among other things told us how the Olympics 2024 were a test of their live subtitling from Danish to Danish, and how they then moved on to the news broadcasting.

We also saw examples of transcription, segmentation and condensation, and how the AI sometimes makes small errors that disturb the meaning of what was said.

Many question arose during the presentation and we had a lively discussion around the use of AI and language.

Navigating Legal and Normative Changes in Technical Communication: AI, Risk Management, and Digital Documentation by Martin Rieder

Martin started out by asking why we need technical documentation, why we're under pressure as technical writers and what the key drivers for change are. He followed up with information on court decisions, social aspect, normative requirements and contractual obligations.

He also reminded us that we must provide the technical documentation in a language that the target group understands - not only in terms of the language itself, but also the words chosen to describe tasks and explain products. In addition, he highlighted the strategic role technical documentation plays in product development and market access processes.

The audience was also informed (and some reminded) of EU legislation as drivers, among others the AI Act and Machinery Regulation, which many of our members work under when creating their documentation.

AI WITH DOCS - FOR REAL NOW! What if AI meets a Help web site? by Kai Weber

Inspired by our president, Kai started his presentation with a stretching exercise to find out, why the participants use AI in their documentation. Surprisingly, most do it, because they want to, and not because they must do it.

He talked about their help site at SimCorp and the implementation of a chatbot incl. expectations, results and learnings.

Kai showed examples of how they had improved and updated the metadata to support the chat bot in finding the right answers.

SURPRISE EXTRA PRESENTATION

Michael Flaskgaard from Siemens Garmesa felt inspired and wanted to share his experience in doing prompting. He freestyled a short presentation on their guidelines and experiences from the past 6 months using Co-Pilot for different use scenarios.