Ask a public company today where artificial intelligence touches its customers and you will receive an answer crafted to be simultaneously impressive and uninformative: AI is embedded across our operations, driving efficiency and innovation.
This vagueness is a choice, and it is aging badly.
Customers increasingly want to know when they are talking to software. Regulators in several states already require it in narrow contexts. Plaintiffs' lawyers have discovered that undisclosed automation makes an excellent exhibit. The direction of travel is not mysterious, and companies face a familiar decision: define the disclosure norm voluntarily, or have it defined for them, adversarially, one enforcement action at a time.
The model already exists. Companies disclose their auditors, their material risks and their executive pay not because the information flatters them but because standardized disclosure converts a suspicion into a fact. A short, plain statement of where automated systems make or shape decisions affecting customers, in service, in pricing, in claims, in hiring, would cost little and preempt much.
The objection is always competitive sensitivity, and it is mostly theater. Rivals do not learn trade secrets from a sentence saying customer service inquiries are initially handled by an automated system with human escalation. What companies are actually protecting is the ambiguity itself, the freedom to let customers assume more human involvement than exists.
That ambiguity was an asset. It is becoming an exposure. The firms that disclose first will look confident. The firms that disclose under subpoena will look like they had something to hide, because by then, legally speaking, they did.
Earlier Cranberry Journal coverage examined Private Companies Reconsider How They Measure Executive Visibility and Small Models, Big Deployments.
