Permission to be random

This is an interesting article on Clorox using genAI in various departments. One part stands out to me:

When it came to generating ideas, it seems the bot's greatest value wasn't so much that it was AI ... it's that it had freedom to propose ideas without fear of scorn or ridicule:

  • People could quickly and openly reject an idea that comes out of the bot without fear of hurting its feelings. It has no feelings! (As noted in the article: most of the bot's ideas were terrible and were tossed out.)
  • The bot didn't have to hold back for fear of being fired. It can't be fired! Not in the traditional sense, at least.
  • It was able to draw odd connections that many humans would have trouble making. (This is not unlike the famous "evolved antenna" example from genetic algorithms.)

In essence: the bot created an environment where weird, wild ideas could come and go without fear.

We often talk about "out of the box" ideas and, well, you get those by having the freedom to go outside of the box.

For your next brainstorming exercise, it might be interesting to mix anonymous ideas from people with ideas from a bot. People can later claim credit for their good ideas, and everyone can quietly blame the bot on anything terrible.

A very human problem

Periodic reminder: do not let the machines run unattended.

The right tool for the job

Sometimes – oftentimes – that genAI bot is overkill