AI Usage Disclosure

Where AI is used

Cultiv.ai uses AI in two places: the Ask Cultiv chat assistant on this website, and as a teaching tool inside CultivLab cohorts. This page covers both.

Ask Cultiv — the website chat assistant

Ask Cultiv is a constrained admissions and program assistant. It is powered by Anthropic Claude (Haiku 4.5 for routing and checks; Sonnet 4.6 for parent-facing answers). It identifies itself as “AI guide · founder-reviewed.”

Ask Cultiv is permitted to:

• Answer questions about CultivLab programs, format, and free beta status
• Explain how parent-led registration works
• Describe how student data is handled
• Suggest a program track based on your student's interests
• Help you book a call with Ash

Ask Cultiv will never:

• Invent pricing, dates, instructor names, or program guarantees
• Give medical, psychological, legal, or college admissions advice
• Ask for your student's last name, phone number, or personal details
• Claim to be a person, parent, or licensed professional
• Answer questions outside the scope of the Cultiv program

Answers mentioning sensitive topics (pricing, safety, legal, student data, college admissions) are automatically flagged for review by Ash within 24 hours. This is Founder Review Mode — every flagged session is read by a human.

AI inside CultivLab cohorts

Students in CultivLab cohorts use AI tools as building instruments — not as answer machines. Mentors teach students when to use AI, when to question it, and how to keep their own voice in their work. The specific tools used vary by cohort track; parents receive the tool list before the cohort begins. We use free tiers of consumer AI tools where possible.

Model training

Your input is not used to train AI models. Anthropic's API does not train on customer inputs by default (Anthropic Trust Center). Cultiv.ai does not share conversation data with any other AI provider.

LLM provider

We use Anthropic (Claude Haiku 4.5 and Claude Sonnet 4.6) for all AI features. Any change of provider will be noted in our decision log and this disclosure will be updated.

Limitations and errors

AI systems can make mistakes. Ask Cultiv may occasionally give an inaccurate or incomplete answer. If anything the assistant says seems wrong or concerning, please reach out directly to Ash at privacy@cultiv.ai. All chat sessions are available for Ash to review.