Build an AI Acceptable Use Policy That Employees Actually Follow

Employees may hide their true AI usage (Shadow IT) out of fear of being reprimanded.

Explicitly state that the survey is anonymous and the goal is to safely enable AI use, not punish past behavior.

The 'Do's and Don'ts' become too restrictive, driving employees back to hidden Shadow IT.

For every 'Don't', try to provide an alternative 'Do' (e.g., 'Don't put PII in public ChatGPT. DO use our enterprise-secured AI tool').

The policy becomes outdated within months due to rapid AI advancements.

Treat the playbook as a 'living document' and strictly adhere to the monthly review cadence to update approved tools.