External AI Communications Guide

How to use this document

• What it is for. This is a base guide that sets how [ORGANIZATION NAME] talks about artificial intelligence externally: in marketing, advertising, sales, press, social media, commercial proposals and incident communications. Its central aim is that no AI claim is published without a basis that supports it, and that capabilities, results, autonomy and regulatory standing are not overstated.
• Who should own it. Usually communications or marketing, with legal as a mandatory reviewer of the highest-risk claims. Enter the function in [RESPONSIBLE FUNCTION] and the person in [OWNER].
• Decisions to take first. Decide who approves each type of message (Section 7), who keeps the substantiation file for each claim (Section 4), who the authorized spokespeople are (Section 8) and whether the organization allows comparative claims about competitors (Section 3).
• Fields to complete. Replace the bracketed text, such as [ORGANIZATION NAME], [RESPONSIBLE FUNCTION], [AI COMMITTEE] and [COMPLETE]. Complete the table in Section 5 with real phrases your team has used or proposed.
• Country variants. This document has no country variants. Rules on misleading advertising, synthetic content labeling, use of people's likeness or voice and mandatory notices vary by country and industry. Use "Related regulatory references" to identify those that apply and adjust Sections 4, 6 and 11.
• Review. Ask counsel to review the adapted version, especially Sections 3 to 6, and to validate the first claims approved under the Section 7 procedure.
• Approval and publication. Have it approved by management or the [AI COMMITTEE], record the approval in "Version history" and distribute it to marketing, communications, sales, talent acquisition, customer service, investor relations and outside agencies. It is an internal document; it is not published.

External AI Communications Guide

Field	Value
Organization	[LEGAL ENTITY NAME]
Owning function	[RESPONSIBLE FUNCTION]
Guide owner	[OWNER]
Mandatory legal review	[COMPLETE: legal function or person]
Approved by	[APPROVED BY]
Effective date	[DATE]
Review frequency	[REVIEW FREQUENCY]
Contact for questions	[AI GOVERNANCE EMAIL]

1. Purpose and scope

1.1 This guide sets the rules under which [ORGANIZATION NAME] communicates its use of artificial intelligence in its products, services and operations to customers, users, media, investors, candidates, authorities and the general public.

1.2 It applies to all staff and to agencies, consultants, business partners and influencers communicating on behalf of the organization, and to every format: website and product pages, ads, marketing emails, sales decks, responses to tenders and customer questionnaires, press releases, interviews, social media, case studies, in-product copy, job postings and investor materials.

1.3 It applies together with the Responsible AI Use Policy (Including Generative AI), the Public Statement on Responsible AI Use, if any, the AI Incident Response Plan, the AI Interaction Notice and Synthetic Content Labeling and the organization's brand guidelines. The detailed method for substantiating claims is in the AI Marketing Claims Substantiation Procedure: this guide sets the communication rules and points to it for substantiation files and their register.

2. Principles

All external AI communication by [ORGANIZATION NAME] is:

a) Truthful. It describes what the system does today, not what it will do in the future or what it does under ideal conditions.
b) Substantiated before publication. Every objective claim is backed by sufficient evidence at the time it is published, not afterward.
c) Complete. It does not omit limitations, conditions or human involvement that would change the audience's impression.
d) Proportionate. The level of detail and caution grows with impact: a message about a writing assistant does not need the same as one about a credit decision or diagnostic tool.
e) Consistent. It matches what contracts, technical documentation, the privacy notice and responses to authorities say.
f) Clear. It can be understood without technical knowledge: first the effect on the person, then the mechanism.