Ethics in public relations gets the most attention after something has gone wrong — a statement that omits an important fact, an AI-generated passage containing an error, or an organization that waits too long to correct the record.

By then, the conversation has shifted to damage control, writes Matthew Marcial, CEO of the Public Relations Society of America (PRSA). The better time to raise ethical questions is earlier, while the statement is still being drafted.

Most ethical questions do not announce themselves as dilemmas, Marcial notes. They arrive disguised as routine editing requests: Can we soften this? Do we need to mention that? The reporter didn't ask, so why volunteer it?

The pattern is familiar. A communicator is asked to highlight favorable survey results without noting the small sample size. An executive requests a statement that is technically accurate but omits important context. A team uses generative AI to draft public content without first deciding whether or when they should disclose it.

None of these situations begins as a crisis. Each can become one if employees, reporters or customers later conclude that the organization withheld information or tried to create a misleading impression.

Where the guidance sits

The PRSA Code of Ethics offers a foundation for addressing these moments, according to Marcial. Its professional values include honesty, independence and fairness. Its provisions call on practitioners to preserve the integrity of communication, investigate the accuracy of information released on an organization's behalf, and promptly correct errors.

Those principles belong in the everyday work of media relations, employee communications, executive counsel and content creation — not just in an ethics training session or crisis response, Marcial argues.

In media relations, accuracy involves more than giving a narrowly correct answer to a reporter's question. Communicators should also consider whether a response provides enough context to be understood. Internal communicators face a similar test when explaining a reorganization or policy change. Employees notice when a carefully worded message avoids the question everyone is asking.

AI adds another layer of risk. A draft can look polished while containing a fabricated statistic or language that tracks too closely with someone else's work. Human review cannot stop at grammar and tone, Marcial writes. Someone must verify sources, question unfamiliar claims and watch for bias. If an error gets through, "the AI wrote it" will not satisfy an audience looking for accountability.

Ethical practice also requires communicators to counsel leadership when a proposed message crosses a line — or edges too close to one. That does not require a lecture. Often, the most effective response is to identify what is missing, explain how the audience is likely to read the message, and offer a more candid alternative.

Five questions before sending

Marcial proposes five questions communicators should ask before releasing information:

1. Is it true?
2. Is it complete enough not to mislead?
3. Who could be affected by what we are leaving out?
4. Would we be comfortable explaining this decision publicly?
5. Are we protecting long-term trust or simply solving today's problem?

These questions will not settle every disagreement, but they can create enough space to reconsider language that is accurate only in the narrowest sense, Marcial writes.

He recommends building the questions into editorial checklists, AI policies, crisis protocols and executive-review procedures — making ethical review part of the assignment rather than an objection raised by one person at the eleventh hour.

Trust is rarely lost through one spectacular failure alone, Marcial concludes. More often, it erodes through a series of decisions that seemed harmless when someone clicked "send." For communications teams weighing how to govern AI-assisted content and executive messaging, that puts the burden on process design — before the draft leaves the building.