AI Usage at ICI

Don’t Feed The AI’s #

Without thinking it through first

A lot of initial focus has been put into how we use AI for our own output; how we write or create content.
Let’s start thinking more about what we are inputting into AI.
Despite current policy, assume the worst. While Zoom AI’s current policy is to not use contents of meeting/chats to train the AI, assume it may change it’s policy retroactively. Be wary of what documents you share into the chat, as that can be saved.
Use Recordings/Transcriptions over AI meeting Assistants.
Don’t use third party assistants that have not been vetted.
Be wary of using documents that have private or PID. Assume everything is public.
Review the documents in this past newsletter about how to disable bots. These things can be crazy aggressive in how they track and follow you in future meetings. Some can get into your calendar etc. Don’t use in within your inbox on hard drives.
People at ICI have access to incredibility sensitive data and PID. Do not use AI or upload to AI anything that isn’t already publicly available.
As a rule anything you use or give AI, will be collected and train AI. Imagine that another person could ask AI for that information.