After working in the compliance space for nearly 20 years, it is clear that a lot of company compliance agendas emulate a Potemkin village.
Potemkin village: A construction, either literal or figurative, that intentionally creates a façade of prosperity, progress, or normalcy to mask an impoverished or dysfunctional reality. Inspired by Grigory Potemkin, who allegedly built fake settlements to impress Empress Catherine II during her trip through Crimea in 1787.
We are often contacted by companies scrambling to prepare for an ISO certification or update their registers for an audit. Likewise, for standards, we hear from companies who urgently need to buy documents to satisfy a customer or inspection, despite a limited understanding of their needs. This mad rush to paper over compliance gaps is still dangerous and far too common.
Modern AI promises to reduce our compliance burden by summarizing laws, regulations, and standards, then mapping them to internal procedures and documents. However, the compliance problem that most companies face is not summarizing the laws, nor mapping them to organizational systems — it is operationalizing them on a day-to-day basis and keeping their human workforce abreast of changing requirements.
Relying on an AI tool or a summary can quickly provide a false sense of security. However, poorly synthesizing a requirement — even oversimplifying the wording — may lead to major repercussions. (Never mind the risk that AI tools creating policies, procedures, records, or other “evidence” of compliance may lead your organization into very problematic situations.)
The main problem with relying too much on AI (or any other outsourced work) is that your organization undermines its own compliance expertise, struggling to understand and implement its requirements on a deep level.
The solution is simple and timeless (the opposite of a Potemkin village) — ensure someone at your company has a deep understanding of the actual requirements and how to apply them. Only then can your company start leveraging AI without creating additional work or risk.
