Most corporate codes of conduct were written for a world without generative AI, and it shows. LRN’s 2026 Code of Conduct Report, which benchmarked more than 1,000 codes of conduct against feedback from 2,000 employees across 15 countries, found that only 9% of the codes it analyzed explicitly address AI or technology ethics at all.
That gap matters because it is opening at the same time trust in the code itself is eroding. LRN found that 63% of employees said they used their code as a workplace resource this year, down from 70% in 2025, and just 66% felt confident they could report misconduct without facing retaliation. A quarter of employees said their code is too long or too difficult to navigate to be useful in the moment a decision actually needs to be made, which is precisely the moment an AI-specific provision would need to hold up.
For HR and compliance teams, the fix LRN’s own researchers point to is not writing a new AI chapter from scratch. It is connecting AI use back to the principles codes already contain, accountability, fairness, transparency and sound judgment, rather than treating generative AI as a separate category that needs its own rulebook. That is a materially smaller lift than most compliance functions assume, and it sidesteps the trap of publishing AI guidance so specific it is obsolete before the next model release.
The bigger risk is timing. As AI governance keeps losing the race inside HR, a code of conduct that stays silent on AI is not a neutral document, it is a gap that employees, and eventually regulators, will notice was always there. Pairing a genuine AI provision with the same manager-level reinforcement that HR already uses to keep other policies alive, rather than a one-time training module, is the difference between a code that gets used and one that gets certified and forgotten, a distinction this publication has also raised in the context of how AI efficiency is quietly costing managers their judgment in other parts of the employee lifecycle.
Source: LRN
