SmartVDR — AI & Acceptable Use Addendum
Version 1.0 · Effective Date: [EFFECTIVE DATE]
DRAFT — requires US-licensed attorney review before publication.
This AI & Acceptable Use Addendum ("Addendum") is incorporated into and forms part of the SmartVDR Terms of Service. Capitalized terms have the meanings given in the Terms. If there is a conflict between this Addendum and the body of the Terms regarding AI Features (including the fair use of storage and processing addressed in Section 7), this Addendum controls.
1. What the AI Features Do
SmartVDR offers automated, machine-learning-based features, including: document categorization and suggested renaming; retrieval-augmented question answering ("Q&A" / chat over your documents); red-flag and issue detection; generated summaries, analyses, and reports; and the optional Train Me feature, which analyzes the Sample Reports you upload to adapt AI Output to your writing style (Terms Section 4.5) (together, "AI Features," and their results, "AI Output"). AI Features generate AI Output from Customer Content and from general-purpose AI models operated by us and by third-party AI providers.
Train Me changes how AI Output reads, not whether it is right. Output shaped by your Style Profile will sound more like you, and may therefore *feel* more authoritative and be easier to accept without scrutiny. It is produced by the same models, with the same failure modes described in Section 3, and it carries exactly the same duty of verification. A confident sentence in your own voice is not a verified one.
2. No Advice; No Professional Relationship
The AI Features and AI Output are informational tools only. They do not constitute, and must not be relied on as, legal, financial, tax, accounting, investment, or other professional advice. Using the Service, receiving AI Output, or communicating with SmartVDR does not create an attorney-client, fiduciary, or other professional relationship between you and SmartVDR. SmartVDR is a software provider, not a law firm, and does not practice law or provide professional services.
3. AI Output May Be Wrong — Your Duty to Verify
Automated systems, including large language models, can and do produce output that is inaccurate, incomplete, outdated, biased, internally inconsistent, or entirely fabricated ("hallucinated"), and can miss, misclassify, or misrank material documents, facts, risks, or red flags. AI Output may vary between runs and is not deterministic.
YOU ARE SOLELY RESPONSIBLE FOR INDEPENDENTLY REVIEWING AND VERIFYING ALL AI OUTPUT AGAINST THE UNDERLYING SOURCE DOCUMENTS AND OTHER AUTHORITATIVE SOURCES BEFORE RELYING ON IT FOR ANY DECISION, ADVICE, FILING, OR TRANSACTION. Do not treat AI Output as a complete or authoritative review, as a substitute for professional judgment, or as a substitute for independent due diligence. AI Features are designed to assist qualified professionals, not to replace them.
Where the Service auto-applies an AI suggestion above a confidence threshold (for example, an AI-suggested filename), that automation is a convenience and remains subject to your review; lower-confidence suggestions are presented for your explicit approval or rejection.
4. No Reliance for High-Stakes Decisions Without Verification
You acknowledge that decisions in transactions and diligence can have significant legal and financial consequences. You agree not to rely on AI Output as the sole basis for any such decision and to apply independent professional review appropriate to the stakes. SmartVDR is not responsible for any decision you make, or transaction you enter, in reliance on AI Output.
5. Third-Party AI Providers
AI Features transmit Customer Content (or derived data such as text and embeddings) to third-party AI providers listed in our Sub-processor List to generate AI Output. Under the API terms applicable to our use of these providers, content and prompts submitted through their APIs are not used to train their generally-available models and are retained, if at all, only transiently for abuse-monitoring purposes (which may include limited automated or human review by the provider for policy enforcement). By using the AI Features you authorize this transmission. If you are subject to confidentiality, privilege, ethical, or regulatory obligations that restrict disclosing content to third-party processors, you are responsible for determining whether your use of the AI Features is permitted, and for obtaining any client or third-party consents required, before uploading such content.
6. Acceptable Use of AI Features
In addition to Section 6 of the Terms, you will not: (a) use AI Features to generate content that is unlawful, infringing, harassing, or that you would not be permitted to create yourself; (b) submit content you lack the right or authority to submit to third-party AI providers; (c) attempt to extract, reverse-engineer, or replicate the models, weights, prompts, or logic behind the AI Features or the SmartVDR learning system; (d) use AI Output to build, train, or improve a competing model or service; (e) present AI Output to third parties in a way that misrepresents it as human-reviewed professional advice when it has not been reviewed; or (f) attempt to defeat safety, rate, or usage controls. AI usage, storage, and processing are subject to the Fair Use provisions in Section 7.
7. Fair Use of Storage and AI Features
Paid plans that are described as offering "unlimited" documents, storage, or usage provide those resources for ordinary, good-faith use of a virtual data room in connection with transactions and diligence. "Unlimited" does not mean unrestricted, and it does not entitle you to consume Service resources — including document storage, bandwidth, and the automated AI processing that each uploaded document and each AI operation triggers — in a manner that is excessive, automated, or disproportionate relative to typical use of your plan by comparable customers, or that threatens the stability, security, integrity, or cost of the Service for us or for other customers. Where we advertise a plan as offering "unlimited" documents, storage, or usage, that claim is expressly qualified by this Section, and we give notice of this qualification at the point of sale.
Without limiting Section 6, the following are outside fair use: (a) uploading, re-uploading, or re-processing documents in bulk primarily to consume AI processing rather than to conduct genuine diligence; (b) using the Service as general-purpose bulk file storage, archival, backup, or content-distribution unrelated to a transaction or diligence matter; (c) automated or scripted ingestion, generation, or downloading at volumes or rates far beyond ordinary human-driven use; (d) sharing or coordinating a single account, subscription, or plan across multiple unrelated organizations, deals, or customers to avoid seat, deal, or plan limits; or (e) any use that materially degrades the Service or imposes AI-processing or infrastructure costs on us that are grossly disproportionate to your subscription.
How we enforce fair use. We monitor aggregate usage for these patterns. Where your usage materially exceeds fair use, our normal practice is to notify you first and work with you to bring usage within range or move you to a more suitable plan. Where usage threatens the Service or other customers, or where good-faith engagement fails, we may, with or without prior notice as the circumstances reasonably require, throttle or rate-limit your AI or upload activity, require an upgrade to an appropriate plan, temporarily suspend the offending activity, or, for continued or egregious abuse, suspend the account in accordance with Section 14 of the Terms (which governs notice for account suspension). We will not impose surprise per-item charges for fair-use overages; our remedies are as described here. This does not affect the AI-operation quotas and optional, pre-disclosed AI credit packs described in your plan and the plan catalog, which apply to metered AI operations such as report generation and analysis; those are not fair-use overage charges.
The specific usage levels that we consider consistent with fair use are operational and may change as models, costs, and typical customer behavior evolve; they are not fixed contractual quotas, and we will apply them reasonably and in good faith. If you expect unusually high volume for a legitimate matter (for example, a very large diligence exercise), contact support@smartvdr.com in advance and we will work with you.
8. No Guaranteed Availability of AI Features
AI Features depend on third-party providers and may be modified, degraded, rate-limited, or discontinued. We do not guarantee the continued availability, performance, or output quality of any AI Feature.
9. Your Acknowledgment
By using the AI Features, you acknowledge that you have read and understood this Addendum, that you understand AI Output may be inaccurate or incomplete, and that you are responsible for verifying it before reliance.
