Ethical AI Disclosure

 

Effective date: 13 November 2025 (Australia/Adelaide)
Owner: Michelle Harpas (“we”, “us”, “our”)
Scope: michelleharpas.com and related pages, programs, communities, replays, checkouts, email/messages, and event environments (the “Services”).
Related policies: Privacy Policy, Cookie Policy, Terms & Conditions, Website Disclaimer, Do Not Sell or Share My Personal Information (where applicable)

 

This Ethical AI Disclosure applies to michelleharpas.com and related services operated by Michelle Harpas (“we”, “us”, “our”). It explains our AI practices and safeguards and is incorporated by reference into our Terms & Conditions and Privacy Policy. Your continued use of the website and services signifies acknowledgment of this Disclosure. In the event of any inconsistency, the Terms & Conditions and Privacy Policy prevail.

 

1) Why this page exists

This page explains how we use—and don’t use—AI, the guardrails we apply, your choices (including opt-out), and how we protect your data.

Bottom line: Coaching is human-led. Coaching is human-led. AI supports admin and organisation only—not coaching, ritual design, or decisions about you.

 

2) Our AI Commitments

  • Human-led coaching: Humans coach; AI never replaces coaching, mentorship, or professional guidance.
  • Consent & choice: You can opt out of AI-assisted features involving your data at any time without losing access to programs or support.
  • Data minimisation: We share the least data necessary and redact identifiers (names, emails, locations, minors’ info) where possible.
  • No model training (where offered): We opt out of vendors using your content to train general models whenever that option is provided.
  • No automated decisions. We do not use automated decision-making or profiling that produces legal or similarly significant effects (for example, decisions about eligibility, pricing, or access).
  • Safety first: All practices are voluntary. You can choose alternatives, ask questions, or opt out at any time.
  • Transparency: We document where AI is used and keep this page current. Material changes will be noted in the Change Log.

 

3) Where we use AI (current use cases)

We use AI behind the scenes for low-risk admin only:

  • Summaries & checklists
    • Converting call/video transcripts into session notes, key points, or action lists.
    • Preparing brief meeting summaries for you and/or our internal records.
    • A human reviews AI outputs before they are shared and/or added to program record
  • Content organisation & retrieval
    • Tagging, filing, and searching internal resources (lesson notes, libraries, transcripts)
    • We aim to avoid including personal data in these workflows; where necessary, we redact or anonymise identifiers.
  • Research & synthesis
    • Summarising public information, comparing ideas, and drafting outlines or FAQs.
    • We do not bypass paywalls or ignore robots.txt.
    • A human reviews research outputs before relying on them.
  • Admin drafts
    • Drafting reminders, topic roundups, document structures, and internal naming conventions for team review and approval only.


We do NOT use AI for:

  • Coaching, mentorship, or “ghost-coaching” (impersonating a coach).
  • Diagnosis, treatment, risk scoring, or any kind of clinical or medical judgment.
  • Identity or personality profiling, assigning labels, or inferring protected characteristics.
  • Eligibility or risk scoring for services, discounts, or payment plans.
  • Automated nudges that target vulnerabilities (e.g., pushing purchases based on distress).
  • Designing or scripting rituals, ceremonies, or shamanic/guided journeys (these are crafted by humans).
  • Automated decisions about access, prices, or guaranteed outcomes.

 

4) Your choices (consent, opt-out, preferences)

4.1 AI Opt-Out.
You can opt out of AI-assisted handling of your data at any time. Email support@michelleharpas.com or hello@michelleharpas.com with subject line: “AI Opt-Out”. We’ll confirm and exclude your data from AI-assisted workflows going forward (for example, your transcripts will not be used in AI summarisation).

4.2 Recordings & replays.

  • We announce at the start of a call if it is being recorded or transcribed (e.g., on Zoom).
  • You may keep your camera and/or microphone off and use chat instead.
  • Where feasible, you may request an alternative to a recorded session (see our Terms & Conditions and Website Disclaimer).
  • Remaining in a recorded session after the announcement constitutes your consent to that recording and any stated uses.

4.3 Cookies & ads.

  • Analytics and Advertising/Retargeting cookies run only with your consent.
  • You can manage these via “Manage cookie preferences” on our site.
  • We honour Global Privacy Control (GPC) signals where legally recognised.


No penalty for opting out.

Opting out of AI-assisted processes does not limit your access to coaching, core materials, or support. Some convenience features (such as automated summaries) may simply not be available.

 

5) Lawful bases & regional rights (how we ground this)

We align our AI use with our broader data protection obligations. See our Privacy Policy for full details.

  • Australia (Privacy Act). We apply data minimisation and security safeguards; you may request access and correction and raise complaints through the channels described in the Privacy Policy.
  • EEA/UK (GDPR/UK-GDPR). Depending on the context, we rely on Contract, Consent, Legitimate Interests, or Legal Obligation as lawful bases for processing. You have rights to access, rectification, erasure, restriction, portability, and objection (including to processing based on legitimate interests and direct marketing).
  • California & similar US laws (CPRA et al.). We do not sell personal information. If any cookies or pixels might be considered “sharing” for cross-context behavioural advertising, you may opt out via cookie preferences, GPC, or our Do Not Sell or Share My Personal Information page.

 

6) Data handling & retention (how we protect you)

6.1 Minimisation & redaction.

  • We restrict AI inputs to what is necessary for the specific administrative purpose.
  • Where possible, we mask or remove identifiers (e.g., names, email addresses, locations, minors’ information, therapy history).

6.2 Retention defaults.

  • Raw call transcripts: Deleted within 90 days (or sooner on request), unless law or a specific program requirement requires a different timeframe.
  • Summaries/notes/metadata: May be retained in your program record for continuity and legitimate business/legal purposes unless you request deletion (subject to legal requirements).
  • Analytics and cookies: Stored according to lifespan ranges listed in our Cookie Policy (typically 6–13 months for analytics/ads).

6.3 International transfers.

  • Some vendors process data outside Australia (for example, in the EU/UK and US).
  • We use appropriate safeguards such as Standard Contractual Clauses (SCCs), vendor security controls, and access limitations, as described in our Privacy Policy.

6.4 Security.

  • Admin-only access to AI tools.
  • Least-privilege, role-based permissions and strong authentication.
  • Encryption in transit and vendor data processing agreements (DPAs) where available.
  • Incident response procedures that include AI-assisted workflows.

 

7) Vendors & roles (transparency table)

These are typical vendors currently involved in AI-adjacent workflows. See the Privacy and Cookie Policies for the broader vendor list.

  • Zoom – Live sessions; optional cloud/local recording and transcripts.
    • Role: Processor
    • Training: We do not authorise use of your recordings for model training.
    • Retention: Transcripts and recordings aligned with program needs and the 90-day default, unless law/program terms differ.
  • Paperbell (+ Stripe/PayPal) – Scheduling, client portal, contracts, checkout.
    • Role: Processor
    • Training: N/A (no AI coaching).
    • Retention: Per our Privacy Policy and Paperbell’s own policies.
  • ThriveCart (+ Stripe/PayPal) – Checkout/LMS; minimal necessary cookies.
    • Role: Processor / independent controller for some payment functions.
    • Training: N/A for AI coaching.
    • Retention: Per Privacy Policy and statutory obligations.
  • ConvertKit – Email and automations.
    • Role: Processor
    • Training: N/A for client-content training.
    • Retention: Per Privacy Policy (e.g., while subscribed or as required by law).
  • Descript / transcription tool – Audio/video transcription → summaries.
    • Role: Processor
    • Training: We opt out of model training where this setting is available.
    • Retention: Raw transcripts ≤90 days (unless otherwise required); summaries retained per Privacy Policy.
  • Google Workspace / Drive / Analytics (Consent Mode) – Email, documents, storage, analytics.
    • Role: Processor / controller (as applicable)
    • Training: We avoid uploading personal client content to experimental AI features; where unavoidable, we use anonymisation and opt-out settings where available.
    • Retention: Per Privacy Policy; cookies per Cookie Policy.
  • Project Manager / Notes (e.g., Notion) – Organising checklists and notes.
    • Role: Processor
    • Training: We avoid enabling training on your content where possible; if offered, we opt out.
    • Retention: Per Privacy Policy.
  • Research LLMs (e.g., ChatGPT/Claude/Perplexity) – Public-web research and synthesis.
    • Role: Processor
    • Training: We do not input your personal content without consent or robust anonymisation; we opt out of training where possible.
    • Retention: No client data by default; vendor-side caching may apply per their policies.
  • Analytics/Consent tools (e.g., Google Analytics with Consent Mode, Cookie Compliance™ CMP) – Site measurement and consent management.
    • Role: Processor / controller (as applicable)
    • Training: N/A for AI coaching.
    • Retention: Cookie-level, typically 6–13 months, per Cookie Policy.

We periodically review this table and update it if our stack or configurations change.

 

8) Social media & advertising clarity

We run ads and publish content on platforms such as Facebook, Instagram, Pinterest, YouTube, and LinkedIn.

  • We do not use AI to fingerprint you or build cross-site advertising profiles.
  • Any advertising that reaches you through these platforms is handled by their own tools under:
    • your cookie/consent choices on our site (see Cookie Policy), and
    • your preferences on each platform (e.g., ad settings in Meta, Google).

For more details, see our Cookie Policy and Do Not Sell or Share My Personal Information page (where applicable).

 

9) Children & sensitive topics

  • Our Services are for adults (18+). If you believe a minor’s data has been provided to us, contact us so we can act promptly.
  • We avoid processing sensitive health or biometric data with AI tools.
  • Please avoid sharing highly sensitive information in group settings. If you need more privacy, let us know so we can explore safer options.

 

10) Bias, fairness, and cultural respect

  • We do not use AI to infer or assign identity traits (e.g., race, religion, orientation, political views).
  • A human reviews AI summaries or checklists before they’re shared or stored.
  • Secular language is always available, and we honour non-appropriative, consent-first practice in all animist/shamanic-informed work. You can always request neutral terms (for example, “Inner Ally” instead of “Power Animal”).

 

11) Security, access, and incidents

  • Security: Admin-only access, least-privilege permissions, strong authentication, encryption in transit, and vendor DPAs where available.
  • AI Incident Log: We maintain an internal log for any AI-related issues (e.g., misrouting, over-collection, or mislabelling of data).
  • Notification: If an incident involving AI meaningfully affects your data, we will notify you in line with applicable law and our Privacy Policy.

To report a concern, email support@michelleharpas.com with “AI Security” in the subject line.

 

12) Future AI features (change management)

Before adding new AI features we will:

  1. Define purpose, data scope, lawful basis, and risks;
  2. Keep a human in the loop (no automated client decisions);
  3. Update this page and the Change Log;
  4. Seek additional consent if personal-data use materially expands;
  5. Keep opt-out easy and honoured.

 

13) Legal & Policy Links

 

14) Contact & Rights

  • Email: support@michelleharpas.com
  • Postal: 19 Calstock Avenue, Edwardstown, SA 5039, Australia

You may request access, correction, deletion, and AI opt-out at any time. See the Privacy Policy for your rights by region.

 

15) Change Log

Date Change
13 Nov 2025 First publication