AI Use and Generated Content Policy

1Your responsibility

You control the instructions, context, source material, approvals, and uses of AI output. Review output for accuracy, safety, bias, legality, rights clearance, and suitability before relying on, sharing, publishing, sending, or implementing it.

2Permitted use

You may use AI features for lawful business, creative, operational, analytical, and educational purposes consistent with your plan, these policies, and rights in the materials supplied. You must maintain meaningful human oversight for consequential actions.

3Prohibited content and conduct

Do not use AI features to facilitate illegal activity; exploit or sexualize minors; create nonconsensual intimate imagery; deceive through impersonation or undisclosed synthetic identity; harass or threaten; generate malware or credential theft; evade security controls; infringe rights; or produce content prohibited by the Acceptable Use Policy.

4High-impact decisions

Do not use AI output as the sole basis for decisions affecting employment, housing, credit, insurance, healthcare, education admission, legal rights, or access to essential services. Qualified human review and legally required notice, explanation, testing, and appeal processes must be used.

5Professional and regulated information

AI output is not a substitute for legal, medical, financial, accounting, tax, safety, or other regulated professional advice. Users offering regulated services remain responsible for qualifications, professional standards, supervision, records, and disclosures.

6Inputs, outputs, and ownership

You retain rights you have in submitted material. As between you and FocalShift, and to the extent permitted by law, you may use generated output, subject to third-party rights, connected-provider terms, and these policies. AI output may not be unique, and other users may receive similar output.

7Privacy and confidential information

Do not submit information you are not authorized to process. Use workspace permissions and available privacy controls. Avoid placing highly sensitive data into an AI feature unless the feature is approved for that data and appropriate safeguards, agreements, and lawful bases are in place.

8Disclosure and authenticity

Disclose AI generation or material alteration when required by law, platform rules, professional duties, customer agreements, or when omission would mislead a reasonable person. Do not remove provenance or authenticity signals for a deceptive purpose.

9Automations and agents

Before allowing an AI agent to send messages, publish content, modify records, spend funds, execute code, or take external action, configure authorization, scope, review, limits, and recovery controls. You are responsible for actions approved or enabled through your account.

Human approval is required before an agent action that spends money, moves funds, transfers credits, issues a refund, enters or accepts an agreement, or publishes content to an external audience takes effect. Where the Services provide an approval step, budget cap, spending limit, allowlist, or confirmation prompt for those actions, you must keep it enabled; you may not disable, bypass, script around, or auto-confirm a required approval control. Configure spending and publishing limits appropriate to the exposure you are willing to accept.

Allocation of responsibility. An action an agent takes within the authority you granted is your action, and you are responsible for it as if you had taken it yourself — including a payment made, an agreement entered, a message sent, or content published, and including the case where the agent acts wrongly, inaccurately, or beyond your intent while remaining inside the authority and scope you configured. FocalShift’s responsibility is to provide the authorization, scope, limit, logging, approval, and revocation controls the Services describe, and to operate them in good faith; FocalShift is not responsible for the substance of an in-authority agent action. If a control fails to operate as documented, tell us promptly and we will investigate; the Terms of Service govern any resulting claim.

Agent identity. Where an agent communicates on your behalf with a person outside your account, you must not present it as a human being. Do not remove, obscure, or falsify an agent or automation label the Services apply, and disclose automated communication where law, platform rules, or professional duties require it. You remain the sender of record for anything your agent sends.

10Enforcement and changes

We may limit prompts, outputs, models, agents, or accounts to prevent harm, comply with provider requirements, or enforce this Policy. AI capabilities and safeguards evolve, and this Policy may be updated accordingly.

11جهة الاتصال

AI policy questions or reports may be submitted through FocalShift Contact.

12BYOK and model routing

If bring-your-own-key access is available, you must have authority to use the provider account and follow its terms. Provider credentials must be submitted only through approved secure fields and never placed in prompts, public pages, static files, shared messages, or unapproved configuration. Requests may route through supported system providers, approved workspace providers, or fallback policies based on plan, role, availability, safety, cost, and configuration.

How FocalShift handles your provider keys. A key you supply is stored encrypted at rest, is used only to make requests you or your configured features initiate, and is decrypted only in the moment of use. We do not write a full key to logs, analytics, support transcripts, error reports, or exports, and interfaces display only a truncated reference. Access is limited to the smallest number of personnel who need it for security, incident response, or a support request you make. We do not use your key for another account, and we do not use it to access your provider account for any purpose other than serving your requests through the Services.

You may rotate or revoke a key at any time using available controls, and you should rotate immediately if you suspect exposure; revocation stops future use but does not reverse requests already made or charges the provider has already applied to your provider account. FocalShift may suspend, replace, or refuse to use a key that appears compromised, is rejected by the provider, generates abusive or unsafe traffic, or would cause FocalShift to breach a provider agreement, and may rotate FocalShift-held system credentials at any time. If a provider suspends, rate-limits, deprecates a model, changes terms, or terminates access — whether to your account or to FocalShift — affected features may degrade, queue, fall back to another supported provider where your configuration permits, or become unavailable; FocalShift is not responsible for a provider’s decision, and charges you owe the provider on your own key remain between you and that provider.

13Credits, limits, and audit records

AI use may consume plan allowances, FocalShift Credits, provider credits, or usage-based charges. We may apply rate limits, caps, model restrictions, fair-use controls, and safety review. Consistent with the Privacy Policy, we may record routing, usage, token counts, credit activity, prompts, outputs, errors, and policy decisions for service delivery, security, billing, auditing, and troubleshooting.

14Accuracy and professional judgment

AI output may be inaccurate, incomplete, outdated, biased, offensive, unsafe, misleading, or unsuitable. It does not replace legal, financial, medical, safety, editorial, employment, or other professional review. You must verify facts, permissions, claims, citations, calculations, and compliance before acting on or publishing output.