Generative AI Usage Policy

The S3 Agency
Updated: September 1, 2026

Our position

The S3 Agency uses generative AI to move faster, not to think for us. This policy covers how S3 uses generative AI, what we won’t do with it, and how we protect client information. It applies to all S3 employees, contractors, and freelancers.

1. How we use it

S3 may use generative AI for:

Ideation, exploration, iteration, and first drafts
Research, summarization, and analysis
Concept artwork and comps
Retouching human-generated photography and artwork
Text-to-voice for accessibility
Editing and refining sound and video
Analyzing how brands and their competitors appear in AI-powered search and answer engines
Agentic tools running multi-step tasks under human supervision

Agentic tools have a named person at S3 accountable for their output. They don’t publish, send, purchase, or commit S3 or a client to anything without human approval.

2. Human approval

Unless otherwise directed by a client, S3 may utilize AI-assisted material, including copy, visuals, video, voice, and music, in conceptual and finished work. However, all work requires human approval from appropriate agency talent before reaching a client or the public. A few limits aren’t subject to review or approval. S3 will not:

Use AI to create false, deceptive, or misleading content, including deepfakes or other manipulative media.
Generate or materially imitate the voice, likeness, identity, or performance of a real person without appropriate consent or authorization.
Present AI-generated material as human-made where a client, platform, agreement, or the law requires disclosure. We comply with applicable disclosure requirements imposed by law, platforms, and client agreements.

3. Ownership, intellectual property, and accuracy

Copyright protects human authorship. Material generated entirely by AI, without sufficient human creative contribution, may not be eligible for copyright protection and therefore may have limited ownership protections. AI-assisted work may be protectable to the extent it reflects sufficient human authorship through creative direction, selection, arrangement, editing, modification, or other meaningful human contribution. Where a specific asset is substantially AI-generated, such as an image, voice track, or music bed, its ownership or exclusivity may be limited. We flag material limitations when appropriate so the client can weigh them against how the asset will be used. S3 does not knowingly use AI-generated material that infringes third-party intellectual property. Substantially AI-generated assets intended for commercial use receive appropriate human review before use. AI also fabricates, errs, and reflects bias. AI-assisted work is subject to appropriate human review, including factual verification where factual claims are involved. We review outputs for bias, stereotype, and exclusion. We don’t assume any model is neutral.

4. Client transparency

S3 publishes its current policy here so any client or prospect can see how we work with AI.
We comply with client requests that direct S3 not to use generative AI on their work, in whole or in part.

5. Confidentiality and data security

Public and consumer AI tools may retain, process, or use information entered into them in ways that are not appropriate for confidential client information. Their practices and terms can also change over time. S3 employees use AI tools governed by enterprise, business, or other terms that provide appropriate protections for confidential information, including restrictions on the use of customer inputs for model training. When such protections are not in place, confidential information is not entered into the tool. Confidential information includes agency or client names when not already public, as well as non-public product, service, customer, financial, strategic, proprietary, or personally identifiable information. All AI use must comply with applicable data privacy laws, S3’s confidentiality obligations, and relevant client agreements.

6. Training

Everyone at S3 is trained on responsible AI use, including what these tools can do, where they fail, how bias can appear in their outputs, and how confidential information must be handled. We keep that training current as the technology, risks, and best practices change.

7. Experimentation

S3 encourages experimentation within the standards above. New and experimental AI technologies may be explored, but their use on client work is reviewed and approved before implementation.

8. Accountability

S3 employees report any issue, concern, or possible breach involving AI use to a member of agency leadership to assess for appropriate action. Individuals and teams are accountable for the tools they use and the work those tools help produce. Violations of this policy are treated seriously.

9. Review

S3 reviews this policy at least annually and updates it as the technology, law, industry standards, and client expectations change. Questions about this policy: hello@theS3agency.com