Can you lock down the tone of AI responses?

Summary:

Can organisations lock down AI responses so they use a fixed, corporate style and tone of voice?

Content (please ensure you mask any confidential information):

AI agents seem to allow users to add "like a pirate" or "in the style of Dr Seuss" on the end of queries to get a response in pirate speak.

Although this is obviously a bit of fun to play with, AI agents are a formal channel of communication between employer and employee and I think employers will want to lock the responses down to an appropriate communication style and tone of voice.

Worst case scenario, someone asks an AI agent for a summary of the corporate death in service policy in the style of Dr Seuss and then posts it on social media and it goes viral. That's not something I'd want to see happen.