AI-Generated Readings: Ethics and Disclaimers for Spiritual Practitioners 2026

About one in three Americans consults tarot or astrology at least once a year (Pew Research, 2025). A growing portion of those consultations now involve AI somewhere in the chain - whether the practitioner uses ChatGPT to help interpret a spread, generates a natal chart interpretation with Claude, or delivers a fully automated AI reading as a product.

The tools are usable. The ethics and legal exposure around how you use and disclose them are less clear, and the costs of getting it wrong - to reputation, to client trust, and potentially to regulatory exposure - are real.

This guide covers the practical ethics framework and the specific disclaimer language that reduces risk. It does not constitute legal advice. Regulatory guidance in this area is evolving, and some points are marked [VERIFY] where the specific rule may have been updated since this was written.

What AI Can and Cannot Do in a Reading Context

The criticism from experienced practitioners is specific: AI gives "cookbook answers." Lauren Ash, an astrologer, put it directly - "You can't hold a GPT spiritualist responsible for what it tells people to do" (Yahoo Lifestyle, 2025). ChatGPT can describe the general meaning of Saturn in the 7th house. It cannot hold the complexity of a natal chart where Saturn in the 7th makes aspects to a Moon-Pluto conjunction in the 4th while the client has Scorpio rising and recently ended a 12-year marriage.

That's not a moral argument against using AI. It's a technical argument about what it produces: general frameworks, not individuated readings.

The second concern is about vulnerable users. Research published in ASAP Journal found that AI-spiritual chatbots can promote "delusional thinking" in users who are already in fragile states (ASAP/Review). A client who is in crisis and receives an AI-generated reading as though it's a personalized human interpretation is in a different situation than someone using an AI tool for casual curiosity.

Sources: pbs.org/newshour (PBS NewsHour, 2025); yahoo.com/lifestyle (Yahoo Lifestyle, 2025); asapjournal.com (ASAP/Review)

The Three Categories of AI Use

How you use AI matters for both the ethical and legal analysis. There are three meaningfully different situations:

Category A: AI as a research assistant. You use Claude or ChatGPT to look up the traditional meaning of a card combination, double-check a planetary aspect interpretation, or draft a framework that you then rewrite entirely in your own voice after doing the actual reading. The AI is a reference tool. The reading is yours.

Category B: AI as a draft generator. You feed the chart data or the spread positions to an AI, get a draft interpretation, then edit and personalize it before delivering it to the client. The AI generated the core text; you modified and contextualized it. Disclosure is appropriate here.

Category C: AI as the reading. The client interacts directly with an AI system, or receives an AI-generated output with minimal practitioner review, delivered as a personalized reading. Full disclosure is required.

Most practitioners operate somewhere in Category A or B. Category C is increasingly common as a product category (automated AI reading generators). The disclosure requirements and ethical obligations scale with how much of the reading comes from the AI versus from you.

Regulatory Context

There is no US federal law specifically regulating AI use in spiritual or metaphysical services. That absence doesn't mean no rules apply.

FTC unfair or deceptive practices: The FTC's Section 5 prohibition on deceptive acts applies to representations that are misleading to a reasonable consumer. If a practitioner delivers an AI-generated reading without disclosure and a reasonable client would expect a human reading, this could fall under deceptive practices. [VERIFY current FTC guidance on AI disclosure requirements for 2025-2026, as the FTC has been actively developing AI-specific guidance.]

EEA/UK GDPR and AI processing: If you're collecting a client's birth date, time, and place of birth and feeding that into an AI tool to generate an interpretation, you are processing personal data through a third-party processor (OpenAI, Anthropic). Under GDPR Article 28, this requires a Data Processing Agreement with the AI provider, and your Privacy Notice must describe the AI tool as a processor.

OpenAI and Anthropic both publish DPAs. Confirm the current version at their respective legal pages before using them to process client personal data.

Sources: astroficial.com (Astroficial); gdprlocal.com (GDPR Local)

Required Disclosure Elements

A defensible disclosure covers four things:

1. AI Use Disclosure

State explicitly whether and how AI is used. For Category B:

> "This reading was prepared with assistance from AI tools. The interpretation has been reviewed, edited, and contextualized by [your name], a practicing [astrologer/tarot reader] with [X years] of experience. The final reading represents my professional judgment."

For Category C (automated reading products):

> "This reading is generated by AI based on the information you provided. It is not reviewed by a human practitioner before delivery."

2. Entertainment and Non-Professional Advice Disclaimer

> "This reading is provided for entertainment and personal reflection purposes only. It does not constitute professional psychological, medical, financial, or legal advice. For significant life decisions, please consult qualified professionals in the relevant field."

This language is standard practice in the spiritual services industry. Its function is to establish expectations clearly, not to disclaim responsibility for poor practice.

3. Limitation of Liability

> "[Your business name] is not responsible for decisions made or actions taken based on any reading or interpretation provided through this service."

4. Data Handling Statement

> "Birth data and personal information you provide is used solely to generate your reading. [If applicable: This information is processed through [AI tool name] subject to their privacy policy, available at [link].] We do not sell or share your personal information with third parties. You may request deletion of your data at any time by contacting [email]."

For EU/UK clients, this last section must align with your full Privacy Notice and GDPR obligations.

Where Disclaimers Live

A disclaimer buried in a Terms of Service page that clients never read does not meaningfully change the ethical or legal situation. Effective disclaimer placement:

For automated AI reading products where the client never interacts with you directly, placement at the purchase page is the most critical point. The client needs to understand what they're buying before they pay.

The Downstream Labor Consideration

PBS NewsHour (2025) reported that the growing use of AI for tarot readings is displacing paid practitioners. When experienced readers charge $75-150 for a session and an AI tool delivers a "reading" for $5 or free, the market pressure is real.

This isn't an argument against using AI. It's context for why the disclosure question matters to the community, not just to regulators. Practitioners who use AI transparently and position it appropriately ("AI-assisted" or "AI-enhanced" with human review) differentiate their work from fully automated readings. Those who don't disclose erode client trust in the category broadly.

Source: pbs.org/newshour/arts/more-tarot-readers-are-turning-to-ai-for-advice (PBS NewsHour, 2025)

Practical Workflow for Category B (AI-Assisted Readings)

A workflow that is both practical and defensible:

1. Conduct the actual reading - pull the cards, calculate the chart, observe the aspects.
2. Use AI to generate a draft interpretation of specific elements (the meaning of a planetary configuration, possible themes for a particular card in a particular position).
3. Read the AI output critically. Cross-check against your own knowledge. The AI is frequently generic and occasionally wrong.
4. Rewrite the interpretation in your own voice, incorporating the specific details of the client's question, the overall pattern of the spread or chart, and your professional intuition.
5. Deliver the final reading under your name with appropriate disclosure.

The key discipline: you are responsible for the reading you deliver. AI as a lookup tool or first-draft generator doesn't transfer that responsibility to the model.

What to Avoid

Medical or mental health claims: A reading that suggests a client's natal chart indicates a health condition, or that a tarot spread reveals the source of a mental health struggle, creates serious liability exposure. Stay entirely clear of health outcomes in both the reading and in testimonials you collect.

Predictive claims presented as certain: "Your reading indicates you will leave your job in the next 3 months" is different from "This spread suggests a theme of transition and release in your professional life." The second framing is defensible; the first is not.

Fabricated AI capabilities: Claiming that your AI tool has intuitive or spiritual capabilities it doesn't have - "our AI is trained on 10,000 readings by master astrologers" without that being literally true - is both dishonest and potentially a deceptive practices issue.

Related Resources

For the legal disclaimers covering readings broadly (beyond AI-specific considerations): legal disclaimers for readings.

For protecting client data including birth data under GDPR: protect client data for readings.

For the AI tool comparison specific to astrology and tarot: ChatGPT vs Claude for astrology and AI ethics in divination.

For GDPR cookie and privacy policy setup: GDPR and cookie consent for spiritual businesses.

Frequently Asked Questions

If I use ChatGPT to help write a reading, do I have to disclose it?

There's no universal legal requirement in most jurisdictions right now - but the ethical argument is clear. If a client asks "did you write this yourself?" and the honest answer is "I used AI to draft it and then reviewed it," disclosure at purchase time is the right approach. Beyond ethics, the FTC's general deceptive practices framework applies if non-disclosure creates a misleading impression about what the client is receiving. [VERIFY FTC's 2025-2026 AI disclosure guidance for service providers.]

Can I sell fully automated AI readings as a product?

Yes, with full disclosure. The product needs to be clearly described as AI-generated at the point of sale - not described in terms that suggest human-crafted personalization. An automated birth chart interpretation tool marketed as "your personalized natal chart reading" with no disclosure that it's AI-generated creates deceptive practices risk. "AI-generated birth chart interpretation" is defensible. The client knows what they're buying.

What happens to my client's birth data when I use ChatGPT to generate a reading?

Birth date, time, and place entered into ChatGPT's standard interface are processed by OpenAI subject to their privacy policy. OpenAI has a DPA available for business accounts - if you're using the API or a ChatGPT Teams account, you can process client data under that agreement with appropriate data handling controls. The default consumer ChatGPT interface doesn't offer the same DPA protections. For EU/UK clients with GDPR implications, use the API under a DPA rather than the consumer interface.

Should I get consent before feeding a client's birth data into an AI tool?

For EU/UK clients under GDPR: yes, explicit consent (or a clear contractual basis) is required before processing personal data through a third-party processor like OpenAI. Your Privacy Notice must name AI tools as processors and describe how data is handled. For US clients: no specific law requires it, but notifying clients in your service description is good practice and reduces the risk of surprise and complaint.

What if a client says the AI reading gave them bad advice that harmed them?

This is why disclaimers matter and why they need to be placed where clients see them before purchase. A clear "entertainment purposes only, not professional advice" disclaimer that the client acknowledged at purchase creates a documented record that expectations were set correctly. If a client is in distress and reaches out, respond with empathy and direct them to appropriate professional resources. Keep records of your disclaimer placements. Consult a lawyer if a formal complaint arises.