AI Output Disclaimer

1. Nature of AI Output

1.1 How Our Tools Are Built

HarvestEdgeTools (“the Factory”) builds custom applications and tools using a system of AI agents operating under human oversight. Our AI agents are powered by large language models (primarily Anthropic’s Claude) and perform functions including gathering requirements, generating code, creating content, testing applications, and validating theological accuracy.

1.2 AI-Assisted, Human-Supervised

All output produced by Factory AI agents is AI-assisted and human-supervised. Our AI agents are tools that augment human capabilities; they do not operate autonomously without oversight. However, AI agents may generate output containing errors, inaccuracies, omissions, or limitations not immediately apparent.

1.3 Output Provided “As-Is”

All tools, applications, code, content, and other output produced through Factory services (“AI Output”) are provided on an “AS-IS” and “AS-AVAILABLE” basis. AI Output is inherently probabilistic and may not be error-free, complete, or suitable for every intended use without additional review.

1.4 Partner Review Responsibility

Partners are responsible for reviewing, testing, and validating all AI Output before deploying it in production environments or relying upon it for ministry, business, or any other purpose.

2. No Warranties

2.1 Disclaimer of Warranties

TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE FACTORY DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, WITH RESPECT TO AI OUTPUT, INCLUDING BUT NOT LIMITED TO:

• Warranty of Merchantability: The Factory does not warrant that AI Output is of merchantable quality or fit for commercial use.
• Warranty of Fitness for a Particular Purpose: The Factory does not warrant suitability for any particular purpose.
• Warranty of Non-Infringement: The Factory does not warrant that AI Output does not infringe third-party rights.
• Warranty of Accuracy or Completeness: The Factory does not warrant accuracy, completeness, currency, reliability, or error-free operation.
• Warranty of Theological Correctness: Despite our AQUILA validation system, the Factory does not warrant that theological content is doctrinally correct or pastorally appropriate.
• Warranty of Security: The Factory does not warrant that generated code is free from vulnerabilities or defects.
• Warranty of Uninterrupted Service: The Factory does not warrant availability without interruption.

2.2 No Professional Advice

AI Output does not constitute professional advice of any kind, including legal, financial, tax, accounting, medical, or counseling advice.

2.3 Acknowledgment

By using Factory services, Partners acknowledge that they understand the inherent limitations of AI-generated output and accept the risks associated with its use.

3. AQUILA Theological Validation

3.1 What AQUILA Does

Theological content produced through Factory services is validated against the Baptist Faith & Message 2000 (BFM 2000) as one unified confessional standard. AQUILA treats the BFM 2000’s 18 Articles as a cohesive whole — not a tiered hierarchy — reflecting the SBC’s adoption of the BFM 2000 as a complete statement of faith.

In normal operation, AQUILA uses an AI language model to evaluate content against all 18 BFM 2000 Articles. When that model is unavailable, AQUILA falls back to explicit automated keyword checks on Article I (The Scriptures) and Article IV (Salvation), with limited keyword-based detection for select additional doctrinal areas; deeper per-Article validation logic is under active development. AQUILA’s capabilities and limitations are described further in Section 3.2.

User-facing Scripture references use the Christian Standard Bible (CSB) translation.

3.2 AQUILA Is a Tool, Not a Guarantee

AQUILA is a validation tool, not a guarantee of theological correctness. Despite our commitment to theological rigor:

• AQUILA does not guarantee that all doctrinal errors will be identified and corrected.
• Theological validation is a tool to assist Partners, not a replacement for the discernment of qualified pastoral leadership.
• The volume and nuance of theological content may exceed the capacity for exhaustive review of every doctrinal implication.
• AI systems and human reviewers are both subject to limitations, biases, and the possibility of error.
• AQUILA validates against the BFM 2000 as a baseline; individual churches may hold additional or more specific doctrinal positions not addressed by the Factory’s validation system.

3.3 Theological Responsibility

The Partner bears ultimate responsibility for theological content used in its ministry. Use of Factory services does not create a pastoral, counseling, or spiritual direction relationship.

3.4 Denominational Scope

The Factory’s theological validation system is designed around the Baptist Faith and Message 2000. Partners from other Baptist traditions or Christian denominations should exercise additional care in reviewing AI-generated theological content to ensure alignment with their own doctrinal standards.

4. Limitation of Liability

4.1 Liability Cap

TO THE MAXIMUM EXTENT PERMITTED BY LAW, THE TOTAL AGGREGATE LIABILITY OF HARVESTEDGETOOLS, INC. SHALL NOT EXCEED THE TOTAL MONTHLY PARTNERSHIP DONATIONS ACTUALLY PAID BY THE PARTNER DURING THE TWELVE (12) MONTHS PRECEDING THE CLAIM.

For example, for a Partner giving $5.00/month, this cap equals $60.00.

4.2 Exclusion of Consequential Damages

IN NO EVENT SHALL HARVESTEDGETOOLS BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, PUNITIVE, OR EXEMPLARY DAMAGES, INCLUDING BUT NOT LIMITED TO:

• Loss of revenue, income, or profits
• Loss of data or data corruption
• Loss of goodwill or reputation
• Business interruption
• Cost of substitute goods or services
• Any other intangible losses

WHETHER OR NOT THE FACTORY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES, AND REGARDLESS OF THE CAUSE OF ACTION OR THEORY OF LIABILITY.

4.3 Essential Purpose

These limitations apply even if any limited remedy fails of its essential purpose. The donation amounts reflect the allocation of risk in this agreement.

4.4 Exceptions

Nothing limits liability for fraud, intentional misrepresentation, or any liability that cannot be excluded under applicable law.

5. Users Should Review All Output

5.1 Pre-Deployment Review

Partners should review all AI Output before deployment, including:

• Code Review: Check for security vulnerabilities, logic errors, and standards compliance.
• Content Review: Check for accuracy, appropriateness, and theological alignment.
• Testing: Test in staging environments before production deployment.
• Theological Review: Have qualified ministry leaders review all theological content.

5.2 Risk Classification

Partners should be especially diligent for high-risk applications:

• Financial tools: Verify PCI-DSS compliance, calculation accuracy
• Children’s ministry tools: Verify COPPA compliance, parental consent
• Membership data tools: Verify privacy controls, access restrictions
• Public-facing content: Verify accuracy, brand alignment, theological soundness

6. Continuous Improvement

6.1 Ongoing Development

The Factory continuously improves its AI agents, processes, and services. As a result:

• The capabilities, accuracy, and quality of AI Output may change over time.
• Output produced at an earlier date may not reflect current capabilities.
• New features and validation processes may be added, and existing ones may be modified or retired.

6.2 No Obligation to Update Past Output

The Factory is not obligated to retroactively update or improve AI Output that has already been delivered, except as specifically agreed upon in an active project scope.

7. Acceptance

By using Factory services, Partners acknowledge that they have read, understood, and agree to this AI Output Disclaimer. This Disclaimer is incorporated by reference into the Terms of Service.

8. Contact Information

DRAFT — NOT LEGAL ADVICE — REQUIRES ATTORNEY REVIEW