Being interrupted is annoying. Apparently even AI-generated podcast hosts agree.
At least that's what Google NotebookLM users came upon. Launched last 12 months, NotebookLM went viral for its feature that creates fully AI-generated podcast-like discussions from user-uploaded content and is discussed by chatty AI bots that act like podcast hosts. In December 2024, NotebookLM introduced a brand new feature called “Interactive Mode,” which allows the user to “call in” to the podcast and ask questions, effectively interrupting the AI ​​presenters while they speak.
When the feature was first introduced, the AI ​​hosts seemed annoyed by such interruptions. They occasionally gave human callers short comments like “I used to be just there” or “As I used to be about to say,” which felt “oddly controversial,” Josh Woodward, vice chairman of Google Labs, told TechCrunch.
So the NotebookLM team decided that some “usability optimization” was so as and posted a self-deprecating joke about it on the product’s official X account:
Woodward said the team solved the issue partly by studying how its own members would respond more politely to interruptions.
“We tested different prompts and infrequently examined how team members would reply to interruptions. “We got here across a brand new prompt that we expect feels friendlier and more engaging,” he said.
It's not entirely clear why the issue occurred in the primary place. Human podcast hosts sometimes express frustration when interrupted, which will be reflected in a system's training data. However, a source acquainted with the matter said that this case was most definitely resulting from the system's prompt design and never training data.
Regardless, the update appears to be working. When TechCrunch tried out the interactive mode, the AI ​​moderator didn't sound upset, but expressed surprise, shouting “Woah!” before politely asking the human to chime in.
.
