Effective Date: 4/1/2026

Purpose and Scope

This AI Use Policy applies to all team members, contractors, and collaborators of Small Projects Bureau (“SPB”) who use artificial intelligence (AI) tools in the course of their work for the studio. It covers generative AI, analytical AI, and agentic AI used for strategy, branding, design, marketing, operations, and internal productivity.

The purpose of this policy is to:

• Ensure AI is used in a way that advances client outcomes while protecting confidentiality, intellectual property, and trust.
• Provide clear expectations for how AI can and cannot be used in SPB’s work.
• Maintain human judgment, creativity, and accountability at the center of all client-facing deliverables.

All AI use at SPB must follow these core principles:

• Human-led, AI-assisted: AI is an assistive tool, not an autonomous decision-maker. SPB humans remain fully responsible for all strategic recommendations and deliverables.
• Client-first, trust-centric: We protect client confidentiality, honor contracts and NDAs, and avoid AI uses that could compromise client trust or reputation.
• Transparency and honesty: When AI meaningfully contributes to a deliverable or process that might affect client decisions, we are prepared to explain how, where, and why it was used.
• Privacy and compliance by design: We handle data with minimalism (only what’s necessary) and follow emerging state-level AI, privacy, and discrimination laws, particularly where clients or end users reside.
• Bias awareness and equity: We recognize that AI systems can encode bias, and we commit to review, edit, and challenge AI outputs—especially where work touches audiences who have been historically marginalized.

Approved Use Cases

SPB team members may use approved AI tools for low- and moderate-risk tasks, provided that human review is always applied before anything reaches clients or the public. Examples include:

Ideation & brainstorming

• Exploring naming territories, campaign concepts, story angles, and content themes.
• Generating alternative framings for positioning, messaging, and narratives that SPB then refines.

Content drafting and refinement

• First-draft internal documents, such as meeting notes, summaries, task lists, or internal SOPs.
• Early drafts of blog posts, emails, social media concepts, or ad copy, with final language always edited and approved by a human strategist or writer.

Research support

• Quick landscape scans, terminology checks, and synthesizing publicly available information as starting points for deeper, human-led research.
• Drafting summaries of long articles or transcripts to accelerate understanding, with humans validating key facts before use.

Design and visual exploration

• Generating moodboards, style explorations, and visual prompts to explore directions—not to replace original design work.
• Creating conceptual images for internal presentations or early-stage ideation, when that does not infringe copyright or client brand guidelines.

Operational efficiency

• Drafting emails, recaps, and checklists, summarizing meetings, and helping with project management artifacts.
• Automating low-risk workflows like document formatting, tagging, or basic data normalization.

In all cases, AI outputs are starting points, not final products. Every deliverable used externally must be reviewed, edited, and approved by an SPB human.

Prohibited and High-Risk Uses

Certain uses of AI are prohibited or require explicit partner approval and legal review:

• No confidential or sensitive client data in public tools
• Do not paste non-public client data, strategy documents, financial information, access credentials, or any content covered by NDA into public AI tools (e.g., consumer chatbots) unless a specific data protection agreement is in place.
• No autonomous decision-making for people’s access to services
• SPB will not deploy AI to make consequential decisions such as eligibility for services, hiring, firing, compensation, or access to benefits for any client or audience.
• No unreviewed AI-generated facts or legal/financial claims
• AI outputs must not be treated as authoritative sources for facts, legal advice, medical guidance, or financial recommendations. All such content must be independently verified via reliable human-reviewed sources.
• No deceptive or undisclosed synthetic media
• SPB will not use AI to create deepfakes, impersonations, or misleading synthetic audio or video, especially of real people, without explicit informed consent and clear contextualization.
• No hidden surveillance or monitoring
• AI will not be used to secretly monitor employees, contractors, or client staff, or to track behavior beyond what has been transparently disclosed and agreed.

Where work touches higher-risk domains (e.g., employment, housing, health, insurance, lending, or services in jurisdictions with strict AI laws such as Colorado), AI-related workflows must be escalated for review and, if necessary, excluded from those processes.

Data, Privacy, and Security

To protect clients, partners, and SPB:

• Data minimization: Only upload the minimum necessary context into AI tools. Summarize or abstract client scenarios when possible, removing identifying and sensitive details.
• No regulated personal data: Do not enter social security numbers, full financial account numbers, detailed health information, or other highly sensitive personal data into AI systems.
• Tool vetting: Before adopting a new AI tool, SPB will evaluate its security posture, data retention policies, and model training practices (e.g., whether user content is used for model training).
• State-level compliance awareness: SPB acknowledges that there is no unified federal AI standard, and that obligations now fall primarily to states and municipalities, including disclosure, bias audits, and consent rules in certain contexts.

Human Review, Quality, and Accountability

All SPB work that incorporates AI must undergo human review commensurate with its impact, including:

• Human-in-the-loop: Every client-facing deliverable (strategy, copy, design, reporting, recommendations) must be read, checked, and intentionally approved by a qualified SPB team member.
• Fact-checking: Where AI has been used for research or drafting factual statements, key facts must be verified using trusted sources before publication or client delivery.
• Bias and representation review: When AI is used for messaging, personas, visuals, or audience insights, reviewers must check for harmful stereotypes, skewed representation, or exclusionary framing, and correct as needed.
• Accountability: AI cannot be blamed for errors. Humans remain accountable for the accuracy, ethics, and appropriateness of the work delivered under SPB’s name.

Intellectual Property and Attribution

SPB respects intellectual property (IP) and uses AI in ways that minimize infringement risk:

• AI-generated content as input, not direct output: Where IP risk is unclear (e.g., image generation trained on copyrighted datasets), AI outputs should be treated as conceptual references, not as final artwork or logos, unless cleared by counsel and aligned with client risk appetite.
• Client ownership: Project contracts will clarify how AI-assisted work is handled in relation to ownership and licensing, but as a standard, completed deliverables provided to clients are treated as work-for-hire or licensed according to existing agreements.
• Third-party materials: AI tools may suggest text or images that resemble or reproduce third-party content; it is the responsibility of SPB reviewers to avoid copying protected material and to ensure originality and proper licensing.

Training, Literacy, and Continuous Improvement

SPB will build AI literacy within the team and treat this policy as a living document, including:

• Onboarding: New team members will receive a briefing on this AI policy, including approved tools, prohibited behaviors, and escalation paths.
• Ongoing training: SPB will periodically share updates, examples, and best practices for using AI in branding, strategy, and operations—particularly as laws and tools evolve.
• Feedback loop: Team members are encouraged to raise concerns, share edge cases, and propose improvements to this policy so it remains relevant and practical.
• Governance, Exceptions, and Review
• Policy steward: A designated AI policy lead or partner will steward this policy, maintain the list of approved tools, and serve as a point of contact for questions.
• Exceptions: Any exception to this policy must be explicitly approved in writing by an SPB partner, with documented rationale and risk mitigation steps.
• Review cadence: This policy will be reviewed at least annually or whenever there is a significant shift in AI capabilities, client needs, or regulatory requirements, including new state-level AI laws relevant to SPB’s work.

How This Affects Day-to-Day Work

For everyday SPB practice, this policy means:

• You are encouraged to use AI to speed up low-risk work—drafts, summaries, ideas, and internal organization—while staying within data and confidentiality rules.
• You must never ship AI-generated content directly to a client or the public without thoughtful human review and editing.
• When in doubt about a tool, a prompt, or a use case, you pause and ask—escalate to the AI policy lead or a partner.

By following these guidelines, SPB can leverage AI as a force multiplier for creative, strategic, and operational excellence—without compromising the trust, ethics, and craft that define its work with small businesses, startups, and nonprofits.