What to do when false claims about your company spread online

When allegedly false claims about your company spread, preserve the evidence, verify the underlying facts, and identify who needs reliable information. Coordinate the response with the appropriate operational, communications, and legal owners before repeating the claim or attributing intent. Use the smallest accurate statement that serves the affected audience’s decision.

Disinformation is false or misleading information deliberately produced or spread to deceive or cause harm. NIST’s 2024 Generative AI Profile distinguishes it from unintentional misinformation in its discussion of information integrity. A damaging claim is not automatically false, intentional, or AI-generated. The response sequence below is an advisory business method, not a crisis protocol validated by that publication.

Verify what the company is actually denying

Composite scene: You wake at home to messages about an alarming allegation. A draft rebuttal is already circulating internally. It repeats the whole story, identifies a suspected source, and states a confident explanation that nobody has checked with the operating owner.

Pause the draft and separate the assertions. Which event is alleged? Which part is a factual claim? Which part is opinion or interpretation? What can the company confirm now, and which details remain unknown?

Do not treat a critical account as false because it is unwelcome. A genuine service problem can coexist with inaccurate details. Address the verified problem through the proper owner while correcting only what the evidence supports.

Separate content accuracy from intent and identity

Accuracy asks whether an assertion matches the evidence. Intent asks whether deception or harm was deliberate. Identity asks who created or spread the content. Each requires its own basis.

NIST’s profile discusses how generative AI can facilitate false or misleading content. It does not establish that a particular post was generated by AI or that the person sharing it knew it was false. Do not use the possibility of synthetic content as an explanation before verification.

Qualified counsel should guide defamation, privacy, evidence handling, and disclosure questions. Relevant safety, security, employment, or regulated matters need their appropriate specialist response. Ordinary reputation management cannot replace those processes.

The preserve, verify, and respond sequence

1. Preserve relevant evidence through a lawful route

Record the original source, time, visible content, and relevant context using authorized access. Ask counsel and security owners what preservation is required for the actual incident. Avoid accessing private accounts or systems beyond your permission.

Preserve versions where the content changes. A screenshot or forwarded message may omit context, provenance, or the original location. Identify those limits rather than treating the convenient copy as a complete evidential record. Keep access appropriate to the sensitivity and purpose.

2. Break the allegation into checkable assertions

Create a short list of material claims. Assign each to the operational or specialist owner who can check it against original evidence. Mark supported, contradicted, unresolved, and interpretive elements separately.

Do not let the proposed rebuttal dictate the verification. Ask what the evidence shows before deciding what management wants to say. If an internal record conflicts with the preferred account, preserve the conflict and route it appropriately rather than selecting the more reassuring version.

3. Identify the audience and immediate consequence

Ask who needs information to act. Existing customers with a service concern may require a direct operational answer. Employees may need an instruction about inquiries. A wider public response may or may not serve a useful purpose.

Assess immediate protective needs through qualified owners. If the incident involves safety, compromised information, impersonation, or another urgent matter, follow the relevant response process. Do not wait for a polished reputation statement where a proper protective instruction is needed.

4. Approve a bounded factual response

Have the appropriate owners review the exact statement and channel. State verified facts, current action, meaningful uncertainty, and an update route. Avoid unsupported cause, intent, or identity claims.

Do not repeat more of the allegation than the audience needs to understand the correction. Whether and how to describe the disputed content depends on the actual context and counsel’s advice. A concise response should remain accurate without exposing private information or implying a final finding before one exists.

5. Update the record as evidence changes

Return at the promised checkpoint or when a material development warrants an update. Correct an earlier company statement where the evidence requires it through the authorized process. Preserve the distinction between what was known initially and learned later.

Assign responsibility for monitoring the relevant issue and answering operational questions. Avoid having every executive respond independently with a slightly different account. A current verified reference can help the authorized team keep explanations consistent.

Use an assertion ledger before writing a rebuttal

A working ledger can show the assertion, original location, evidence owner, source checked, current finding, limitation, and approved public wording. Keep private source material restricted.

For a hypothetical claim about a service interruption, the ledger might separate whether an interruption occurred, its affected scope, the cause, and what customers need to do. The company may be able to confirm the first two before it knows the cause.

That distinction allows an operational response without inventing a full explanation. It also prevents a denial of one inaccurate detail from implying that no underlying problem occurred. Ask the relevant owner to verify what each sentence actually denies or confirms.

Prepare a holding statement with real placeholders

Proposed wording:

We are aware of [verified description of the issue]. We have confirmed [publishable facts] and are checking [unresolved question]. We are taking [authorized action]. People affected should [verified instruction, if needed]. Our next update will be available through [channel] at [checkpoint]. We will correct the account as evidence develops while protecting private information.

Use only facts and commitments the company can support. If the cause is unknown, omit a causal explanation. If no protective instruction is needed, do not invent one to make the statement appear decisive.

Have counsel and the relevant specialists review the actual text. This example is not a claim that a generic holding statement satisfies the company’s legal or contractual obligations.

Decide whether a public correction serves the audience

A CEO may feel pressure to answer every visible criticism. Ask what the affected audience needs and whether the chosen response provides it. A direct service explanation may be more useful than a public argument about the critic’s tone.

Do not infer that limited visibility makes a claim harmless or that high visibility makes a public CEO response mandatory. Use verified context and the consequence for the relevant people. Qualified communications and legal judgment should inform the actual channel decision.

If a public correction is appropriate, state the supported fact clearly. Avoid claiming that one correction disproves every concern around the incident. If the issue remains under assessment, keep that uncertainty visible.

Manage the information risk inside the company

The 2023 NIST AI Risk Management Framework organizes responsibility around governing, mapping, measuring, and managing risk. Its use here is an analogy for assigning owners and evidence checks, especially where AI-related content is involved. It is not a general corporate crisis-response certification.

Ask staff to distinguish direct observations from forwarded claims. Give them an authorized route to supply relevant evidence and a current instruction about external inquiries. Avoid requiring them to defend a company account that the incident owner has not verified.

Industrial/organizational psychology, the psychology of work and organizations, helps examine how authority and group processes shape information flow. A clear internal channel should support factual correction rather than produce forced agreement with the CEO’s first interpretation.

Protect people while preserving the evidence

The 2020 NIST Privacy Framework addresses risks associated with processing information about people. The response may involve employee, customer, or third-party information. Ask qualified privacy and legal owners how that material should be collected, accessed, retained, and disclosed.

Do not expose a person’s private details in order to demonstrate that the company is right. Do not circulate an unverified suspected identity through informal executive messages. A privacy-safe correction may need to explain an operating fact without sharing the underlying personal record.

The framework is voluntary guidance. It does not settle the actual legal requirements or establish that a chosen disclosure is permitted. Keep those determinations with qualified advisers and the proper authority.

Test three hypothetical outcomes

The review may find the allegation materially false and support a bounded correction. Preserve the evidence and use the approved wording without adding an unverified motive.

It may find that part is true and part is inaccurate. Address the genuine issue, correct the specific error, and avoid a sweeping denial that the mixed evidence cannot support.

It may remain unresolved. State what the company knows, what it is checking, and what affected people should do now. A complete-looking story is not a substitute for a verified one. The response owner should explain why the remaining gap matters and who is working on it.

Learn from the incident without rewriting the initial record

After the immediate response, compare the original assertions, verification sequence, approved statements, and later findings. Identify a specific gap in ownership, evidence access, or communication if the record supports it.

Correct reusable statements and brief the authorized people who answer future questions. Where earlier communications created consequential legal, customer, or employment implications, seek the appropriate professional review of the next action.

The useful outcome is a response grounded in verified facts and the audience’s real need, with privacy, uncertainty, and accountability preserved throughout the process.