t3n Arbeit in Progress

Sparringspartner ChatGPT: Wie viel KI verträgt gute Kommunikation?

33 min

In this episode, political and communication consultant Franzi von Kempes discusses the role of generative AI in communication, particularly in politics and journalism. She argues that while AI can be a useful sparring partner for structuring ideas and drafting texts, it often fails to capture the subtle nuances, personal tone, and authentic Haltung (stance) that are crucial for meaningful public communication. The conversation explores the risks of generic AI-generated texts, the need for transparency and human oversight, and offers practical advice on maintaining trustworthy communication in an age of information overload.

Summarized by Podsumo

🎧 Listen 🎙️ Ask about this episode

✨ Key Takeaways

1

Franzi rarely uses generative AI for writing her newsletter because AI-generated texts lack her personal tone and subtle emotional nuances, often producing generic phrases like the ubiquitous 'three-pointer' or overusing specific expressions (e.g., 'No Shit Sherlock').
2

Effective communication requires knowing both the core message and the target audience—'if you try to reach everyone, you'll end up reaching nearly no one.'
3

The use of AI in political speeches and opinion pieces is particularly problematic because audiences expect a genuine human stance behind such content, which a machine cannot authentically deliver.
4

Franzi suggests a practical approach to coping with information overload: reading curated newsletters, listening to podcasts, browsing libraries/bookstores, and consciously unplugging from social media to regain focus.
5

Instead of 'human in the loop' being removed by AI agents, Franzi believes that for expressing personal Haltung and reacting to current events, human involvement remains essential and will likely always be necessary.

💬 Notable Quotes

"When I give the AI a topic and ask it to turn it into a text, the output never matches what I want—no matter how cleverly I think I've prompted it. The *life* is missing."
"We must remember: what was *bad before* remains bad even with AI. The AI doesn't fix flawed processes; it might even make them worse."
"If you try to *reach everyone*, you end up reaching *almost no one*. That’s the single biggest misstep in communication."