AI Disclaimer

Effective August 31, 2026 · Palim Beta

        AI is an assistant, not an authority: Treat every response as proposed material. Review the wording, facts, sources, rights, privacy impact, and fit with your project before you rely on, apply, export, or publish it.
      

This notice applies whenever Palim prepares, reviews, transforms, or evaluates material through an AI service, including a service you configure yourself. It supplements the Terms of Use and Privacy Notice.

1. Author Control & Change Boundaries

Palim is designed to present AI-assisted changes as reviewable proposals, suggestions, candidates, or staged actions rather than silent authority. The stable unit is a document selection, paragraph, content object, or explicit operation—not a visual line that changes with window width or zoom. You remain responsible for the final manuscript and delivery decision.

Palim identifies AI-assisted proposals in the authoring workflow and records available action receipts and versions for review and recovery. If law, a platform, a publisher, or a client requires disclosure that material was AI-generated or AI-assisted, you are responsible for making the required disclosure. Palim receipts support review but do not by themselves prove legal or policy compliance.

Review and recovery controls reduce accidental changes; they do not prove that an AI result is accurate, lawful, original, private, or suitable for publication.

2. What an AI Action Sends

When you explicitly start an AI action, Palim sends only the instructions, selected text, bounded project context, and limited task metadata needed for that action. The exact content depends on the task and context preview. A Palim account, product telemetry, ordinary saving, or merely opening a project does not by itself send manuscript content to an AI service.

Palim AI may use approved marketplace routes and upstream services that process task content in other regions under varying privacy, training, retention, and security rules; Palim’s online service records do not store raw task content. A service you configure follows its own terms.
Do not send passwords, API keys, access tokens, private identifiers, privileged material, regulated personal data, or confidential manuscripts unless you have authority and accept the active service boundary.
Palim-controlled telemetry excludes prompts and generated text, but the active AI service necessarily receives the task content you choose to send.

3. Accuracy, Citations & Continuity

AI systems can invent facts, quotations, citations, links, names, dates, statistics, or source details; misunderstand instructions; omit qualifications; and introduce contradictions, anachronisms, bias, stereotypes, unsafe material, or continuity errors. Verify important claims against original, current, authoritative sources. A fluent explanation, confidence score, citation-shaped string, or repeated answer is not proof.

4. Copyright, Other Rights & Confidentiality

The ownership, copyrightability, originality, registrability, and commercial usability of AI-assisted material vary by jurisdiction and service terms. Similar output may be generated for others. Palim does not guarantee exclusivity or non-infringement. You must check copyright, trademark, publicity, privacy, confidentiality, contractual, source-license, and attribution obligations before use.

Do not assume an AI provider treats submitted text as confidential. Only submit material you have the right and authority to disclose under the active service's terms.

5. No Professional or High-stakes Advice

AI output in Palim is not legal, medical, financial, tax, employment, compliance, safety, emergency, or other professional advice. Do not use it as the sole basis for a decision that may materially affect a person's rights, health, safety, finances, employment, education, access to services, or legal position. Consult a qualified professional and the controlling primary sources.

6. Availability, Model Change & Third-party Services

AI availability depends on the active service, account access, region, network, provider rules, model availability, usage limits, and current service readiness. Models and provider behavior can change without preserving identical results. A visible control, saved model choice, or successful connection test does not guarantee that a later task will execute.

Palim AI is not currently available for purchase. Developer-configured and authorized test services must not be mistaken for a generally available Palim-managed service.

7. Review, Recovery & Reporting Problems

Before applying a proposal, compare the affected scope, preserve important originals, and use the available preview, receipt, version, reject, and undo paths. Some results may still require manual repair. If an AI action appears to expose private content, bypass review, alter the wrong scope, or produce a safety or rights concern, stop the action and contact contact@nyxevo.com with the smallest non-sensitive reproduction you can provide.