American Diversity Initiative
AI Use Policy
Effective Date: June 15, 2026
Last Reviewed: June 15, 2026
Business Location: Ohio, USA
Applies To: ADI team members, contractors, content partners, technology partners, administrators, and approved vendors
1. Purpose
American Diversity Initiative uses artificial intelligence to support practical, affordable, and accessible online workplace training.
AI may help ADI improve course administration, learner support, exam review, compliance confirmation, content organization, search visibility, internal workflows, and business operations.
This policy sets clear rules for how AI may and may not be used.
The goal is simple:
Use AI to improve consistency, efficiency, and learner experience while protecting privacy, accuracy, fairness, and trust.
2. Scope
This policy applies to all AI tools used by ADI, including:
• Generative AI tools such as ChatGPT, Claude, Gemini, Copilot, or similar tools
• AI tools built into the ADI learning management system
• AI tools used for exam scoring, exam review, certificate support, learner verification, or compliance confirmation
• AI tools used for website content, sales material, proposals, social media, reporting, analytics, or administration
• AI tools used by employees, contractors, partners, or vendors on behalf of ADI
This policy applies whether the tool is free, paid, embedded in software, or used through a third-party platform.
3. ADI’s Position on AI
ADI uses AI as a support tool.
AI does not replace human judgment, subject matter expertise, legal review, content approval, or final business decisions.
AI may assist with:
• Organizing information
• Drafting internal documents
• Supporting learner administration
• Reviewing exam data
• Flagging possible inconsistencies
• Improving content structure
• Supporting search and website visibility
• Preparing draft sales or partnership materials
• Summarizing internal workflows
• Supporting compliance documentation
AI must not be treated as a final authority.
4. Approved AI Use Cases
ADI may use AI for the following purposes when the rules in this policy are followed.
4.1 Online Learning Administration
AI may support:
• Learner onboarding workflows
• Course recommendation logic
• Learning pathway organization
• Employer account setup support
• Administrative summaries
• Internal LMS support tickets
• Learner progress summaries
• Course completion tracking
AI may not override official LMS records unless reviewed and approved by an authorized ADI administrator.
4.2 Exam Review and Compliance Confirmation
ADI may use AI to support exam-related workflows, including:
• Reviewing exam responses
• Supporting automated scoring where appropriate
• Flagging incomplete answers
• Identifying inconsistent or unusual response patterns
• Supporting certificate eligibility review
• Confirming whether required learning steps were completed
• Creating administrative summaries for employer reporting
• Supporting compliance documentation
AI may help confirm whether a learner appears to have completed the required course and exam process.
However, AI must not be the only decision-maker when:
• A learner may fail an exam
• A learner may be denied a certificate
• A learner’s completion record is questioned
• A learner’s conduct is flagged as suspicious
• An employer requests clarification
• An accessibility accommodation may be involved
• The outcome could affect employment, volunteer participation, school participation, or formal compliance status
In these cases, an authorized human reviewer must review the record before a final decision is made.
4.3 Course Content Support
AI may support:
• Drafting course outlines
• Creating plain-language summaries
• Drafting quiz questions
• Improving accessibility of language
• Organizing learning modules
• Creating draft FAQ content
• Supporting course descriptions
• Repurposing approved content into marketing or learner support material
All course content must be reviewed and approved by a qualified human reviewer before use.
AI must not create final course content, compliance claims, legal statements, or subject matter guidance without human review.
4.4 Website, Search, and AI Visibility
AI may support:
• FAQ development
• Website page drafts
• Course description drafts
• Topic pages
• Search engine optimization
• Answer engine optimization
• Metadata
• Blog outlines
• Social media drafts
All public content must be reviewed before publication.
ADI must not publish AI-generated claims about learner numbers, client results, testimonials, legal compliance, certifications, partnerships, or industry benchmarks unless those claims are verified.
4.5 Sales, Proposals, and Partnership Support
AI may support:
• Discovery call summaries
• Draft proposal language
• Partner research
• Outreach drafts
• Follow-up emails
• CRM notes
• Meeting preparation
• Internal sales process documentation
AI must not approve pricing, contracts, partnership terms, commission structures, or final proposals.
Those decisions remain human-only.
5. Human Review Requirements
Human review is required before AI-supported work is used for:
• Course content
• Exam outcome disputes
• Certificate denial
• Compliance claims
• Website content
• Sales materials
• Social media
• Employer-facing reports
• Public statistics
• Mission or vision statements
• Partner communications
• Legal language
• Privacy-related language
• Accessibility claims
• Harassment prevention claims
• Bias awareness claims
• Workplace compliance claims
Human review means an authorized person checks the output for:
• Accuracy
• Tone
• Completeness
• Privacy risk
• Accessibility
• Fairness
• Brand fit
• Legal or compliance risk
• Alignment with ADI’s approved content
6. Human-Only Decisions
AI must not make final decisions about:
• Course approval
• Content approval
• Legal review
• Financial decisions
• Partner approval
• Contract approval
• Final learner certification disputes
• Termination of customer access
• Final compliance determinations
• Refund decisions
• Strategic company decisions
• Public claims about ADI’s results, clients, partnerships, or reach
AI may support these decisions by organizing information or preparing drafts.
The final decision must remain with an authorized human decision-maker.
7. Data Protection Rules
ADI team members, contractors, and partners must protect confidential and personal information when using AI.
The following information must not be entered into public AI tools:
• Customer data
• Learner names
• Learner email addresses
• Employer account data
• Exam records connected to an identifiable person
• Financial data
• Contracts
• Employee information
• Partner agreements
• Commission structures
• Private business records
• Unpublished course materials unless approved
• Confidential client or customer lists
Where AI is used with learner, employer, or exam-related data, ADI must use approved systems with appropriate privacy, security, access control, and vendor review.
8. Learner Data and Exam Records
Learner data must be handled carefully.
When AI is used to support exam review or compliance confirmation, ADI must limit the information used to what is necessary for that purpose.
Where possible, ADI should use:
• Learner ID numbers instead of names
• De-identified exam records
• Aggregated reporting
• Role-based access
• Secure LMS records
• Approved vendor tools
• Human review for exceptions
AI-assisted exam review must be documented when it affects certificate eligibility, employer reporting, or compliance confirmation.
9. Transparency
ADI should be clear when AI is used in a meaningful way to support online learning, exam review, learner administration, or compliance confirmation.
ADI does not need to disclose every internal administrative use of AI.
However, ADI should disclose AI use when it materially supports:
• Exam scoring
• Exam review
• Certificate eligibility
• Learner progress review
• Employer compliance reporting
• Learner support decisions
Suggested disclosure language:
“ADI may use approved technology, including AI-supported tools, to assist with course administration, exam review, learner progress tracking, and compliance confirmation. AI is used to support consistency and efficiency. Human review is required for disputed outcomes, certificate concerns, accessibility-related issues, or decisions that may materially affect a learner.”
10. Accuracy and No Hallucinated Claims
ADI must not use AI to invent or exaggerate:
• Learner counts
• Client counts
• Testimonials
• Reviews
• Case studies
• Partnerships
• Awards
• Certifications
• Legal compliance guarantees
• Course effectiveness claims
• Industry statistics
• Benchmarking claims
Any claim used publicly must be supported by approved internal records or verified sources.
If a claim cannot be verified, it must be removed or clearly marked as a draft placeholder.
11. Compliance Claims
ADI provides online workplace training to support respectful workplaces, workplace communication, accessibility, inclusion, bias awareness, harassment prevention, and related workplace learning needs.
AI must not be used to claim that a course guarantees legal compliance unless that claim has been reviewed and approved by legal counsel.
Acceptable language includes:
• “Supports compliance efforts”
• “Helps organizations document training completion”
• “Provides practical workplace training”
• “Supports respectful workplace learning”
• “Helps employers provide consistent training access”
Avoid language such as:
• “Guarantees compliance”
• “Eliminates legal risk”
• “Fully protects employers”
• “Automatically certifies legal compliance”
• “Meets every state requirement”
12. Accessibility and Fairness
AI-supported learning and exam tools must be used in a way that respects accessibility and fairness.
ADI must provide human review when:
• A learner requests an accommodation
• An exam response may reflect language, disability, neurodiversity, translation, or accessibility-related needs
• AI flags a learner response as incomplete, inconsistent, or suspicious
• A learner disputes an AI-supported result
• An employer asks for clarification about a learner outcome
AI must not be used to penalize learners without review.
13. Approved Tools
ADI may only use AI tools that have been reviewed for appropriate business use.
Before using a new AI tool, ADI should review:
• What data the tool collects
• Whether data is used to train models
• Where data is stored
• Who can access the data
• Whether the tool has enterprise or privacy settings
• Whether learner or employer information will be processed
• Whether the tool supports audit trails
• Whether the tool integrates with the LMS
• Whether the vendor terms are acceptable
Free public AI tools may be used for general drafting and brainstorming only.
They must not be used with confidential, personal, learner, customer, contract, financial, or employer account information.
14. Vendor and LMS Requirements
When AI is built into the LMS or a vendor system, ADI should confirm:
• The vendor’s privacy terms
• Data storage location
• Security controls
• Access controls
• Audit logs
• Data retention rules
• Data deletion process
• Whether learner data is used for model training
• Whether human override is available
• Whether AI decisions can be reviewed
• Whether records can be exported if needed
ADI should keep vendor documentation on file.
15. Employee and Contractor Responsibilities
Anyone using AI on behalf of ADI must:
• Follow this policy
• Use approved tools
• Protect confidential information
• Review AI output before use
• Fact-check claims
• Avoid uploading restricted data
• Keep ADI’s brand voice clear, practical, calm, and helpful
• Escalate concerns quickly
• Ask for approval when unsure
Failure to follow this policy may result in removal of AI access, contract review, or other corrective action.
16. Recordkeeping
ADI should keep records of AI-supported workflows where the output affects:
• Course completion
• Exam review
• Certificate eligibility
• Employer reporting
• Compliance confirmation
• Public claims
• Customer proposals
• Partnership agreements
• Legal or privacy-related decisions
Records may include:
• Tool used
• Date used
• Purpose
• Human reviewer
• Final decision
• Any edits made
• Any issue or dispute raised
17. Escalation
Team members must escalate AI-related concerns when:
• AI produces inaccurate or harmful content
• AI creates an unverified claim
• AI flags a learner for review
• A learner disputes an AI-supported outcome
• Confidential information may have been entered into an AI tool
• A vendor tool creates privacy concerns
• A customer asks how AI was used
• A compliance-related claim needs review
Escalation should go to:
Primary Contact: lscaravelli@americandiversityinitiative.com
18. Breach or Misuse Response
If confidential, learner, customer, employee, or financial information is accidentally entered into an unauthorized AI tool, the person who entered it must report the issue immediately.
ADI will review:
• What information was entered
• Which tool was used
• Whether the data can be deleted
• Whether the vendor must be contacted
• Whether customers, learners, employers, or regulators must be notified
• What corrective action is needed
19. Brand Voice and Public Communication
AI-generated content must reflect ADI’s brand voice.
ADI’s preferred voice is:
• Clear
• Practical
• Calm
• Helpful
• Respectful
• Actionable
ADI should avoid:
• Hype
• Fear-based messaging
• Political framing
• Aggressive sales language
• Heavy jargon
• Overpromising
• Unverified proof claims
ADI’s public content should focus on practical workplace training, respectful communication, workplace culture, accessibility, inclusion, compliance support, and risk reduction.
20. Review Cycle
This policy should be reviewed at least every 6 months or sooner if:
• ADI adopts a new AI tool
• AI is added to the LMS
• AI is used for exam review or certificate decisions
• Privacy laws or customer requirements change
• A vendor changes its AI terms
• A learner or employer raises a concern
• ADI expands into new jurisdictions or regulated training categories
21. Policy Acknowledgement
All team members, contractors, and approved partners using AI on behalf of ADI must confirm they have read, understood, and agree to follow this policy.