AI Use Disclosure

How AI Works in Behold

Behold Scripture includes optional AI-powered features designed to assist users with Bible study, reflection, and educational exploration.

These features may:

• Answer questions about passages and theological concepts
• Summarize chapters or themes
• Explain historical context, original languages, and cross-references
• Engage your personal notes to deepen study over time
• Offer different study modes (Tutor, Expositor, Skeptic, Pastoral) that change the AI's posture

AI Provider

Behold's AI features are powered by Anthropic's Claude. Your messages and related context (the passage you're studying, your relevant notes) are sent to Anthropic via their API to generate responses. Anthropic processes this content according to their Privacy Policy.

Per Anthropic's API terms, content submitted via the Claude API is not used to train their public AI models.

We may add or change AI providers in the future. If we do, this disclosure will be updated.

What Information Is Sent to the AI

When you use Ask Behold, the following may be sent to the AI provider to generate a response:

• The question or message you typed
• Your recent AI conversation history (so the AI can follow context)
• The passage you're currently studying
• Relevant notes you've created (to personalize responses)

We send the minimum information necessary to produce a useful answer.

AI Limitations

AI systems are probabilistic and have real limitations. AI responses may:

• Contain factual inaccuracies
• Misunderstand context or your question
• Be incomplete or partial
• Reflect interpretations that don't match your specific theological tradition
• Generate content that sounds confident but is wrong

AI responses are not:

• Official doctrine of any specific church or tradition
• Pastoral counseling from a trained pastor or counselor
• Medical, legal, mental health, or financial advice
• Emergency guidance — if you're in crisis, contact a qualified professional or emergency service

Human Oversight

AI-generated content in Behold is automated. Responses are not manually reviewed before being shown to you. We rely on the AI provider's safety systems and on the design of our prompts to keep responses helpful and appropriate, but we cannot guarantee every response will be accurate or suitable for every reader.

Users should independently verify important information, particularly on contested theological questions, historical claims, and applications to personal life decisions.

Sensitive Topics

Behold's AI is designed to engage hard questions honestly, including topics where Christians have genuinely different views (such as same-sex relationships, women's ordination, divorce and remarriage, abortion, and end-times questions). On these topics, the AI is instructed to present multiple faithful Christian perspectives without picking one.

On topics where Christian ethics is clear and unified (such as racism, dehumanization, abuse, and hatred of others), the AI is instructed to affirm scripture plainly.

On personal disclosure (when a user shares something difficult or vulnerable), the AI is instructed to respond pastorally first — before any theological engagement.

If you encounter an AI response that feels inappropriate, hurtful, or unsafe, please contact us at AdamsDev.Labs@gmail.com. Real user feedback shapes how we tune the system.

AI Safety and Abuse Prevention

We may monitor or restrict abusive usage patterns to protect infrastructure, prevent malicious behavior, maintain service reliability, and comply with legal obligations.

Data Usage

We do not sell AI conversation data. Per Anthropic's API terms, your AI conversation content is not used to train their public AI models.

Changes to AI Systems

AI providers, models, and features may change over time. We may add, modify, or remove AI features as the underlying technology evolves. Material changes will be reflected in this disclosure.

Contact

Questions about AI usage in Behold:

Lamar Adams
Behold Scripture
Email: AdamsDev.Labs@gmail.com