A new report from the ethics and compliance firm LRN looked at nearly 200 codes of conduct from companies around the world. Only 15 percent of them mention AI in any meaningful way. That is up from just 5 percent two years ago, but it still means that 85 percent of company rulebooks say nothing about how employees should or should not use the technology.
This gap would be less alarming if employees were not already using AI on the job. But they are. Separate research from IBM found that only 37 percent of organizations have any AI governance policy in place at all. A survey from BlackFog found that about half of employees use AI tools without getting approval from their employer first, and many of them think that is perfectly fine to do.
So picture the situation inside a typical company right now. Staff are quietly using AI to draft emails, summarize contracts, or analyze customer data. Their employer has not told them what is off limits. The code of conduct, the one document meant to spell out right and wrong at work, is silent on the subject. Employees are left guessing, and guessing is exactly how mistakes and legal exposure happen.
LRN's researchers make a useful point: fixing this does not require writing a brand new AI chapter from scratch. It means connecting AI back to principles companies already have, things like fairness, honesty, and using good judgment. A rule that says do not mislead customers already covers a salesperson using an AI chatbot to generate false claims. The company just has to say so clearly.
There is a second problem sitting underneath this one. LRN also found that while 85 percent of executives say they talk to their teams about the code of conduct, less than half of frontline employees say they actually hear about it from their own managers. Rules that exist only on paper at the top of the company rarely reach the people who need them most.
This is happening while trust in the reporting system itself is slipping. A separate survey by HR Acuity found that 55 percent of employees experienced or witnessed misconduct in 2025, up sharply from 41 percent the year before. At the same time, the share of employees who believe they can report a problem without facing retaliation dropped from 71 percent to 66 percent.
Put those two trends together and the picture is not pretty. Companies are asking employees to follow rules that do not mention their newest daily tool, delivered by managers who often do not pass the message along, inside a system that fewer employees trust to protect them if they speak up.
The fix does not need a task force or outside consultants. It needs someone in HR or legal to sit down, read the existing code of conduct, and ask one plain question: if an employee used AI to do something questionable tomorrow, would this document give them a clear answer? For 85 percent of companies right now, the honest answer is no.