At a glance
This document applies to LabMind at labmind.si and to other products released by Sagaway OÜ, which operates them.
- This policy sets out what you may not use the Service for, and where you need to take an extra step.
- It applies to every user, and to every model, execution location and client, including on-device models on your phone and model endpoints you configure yourself.
- LabMind (研云) can run code, work with files, reach the network and operate web pages as you. Actions the AI takes for you are covered by this policy too, and you are responsible for them.
- Breaking this policy can lead to a request being blocked, features being restricted, or your account being suspended or closed.
This policy forms part of the Terms of Service. If the two conflict, the Terms of Service prevail.
Part 1. What is prohibited for everyone
The following apply to all users and all uses.
1. Do not break the law or help others break it
- Do not submit or generate unlawful content, or use the Service to carry out or assist unlawful activity.
- Do not infringe anyone’s intellectual property.
- Do not breach applicable export control or sanctions rules.
2. Do not endanger critical infrastructure
Do not use the Service to attack or disrupt critical systems such as power, water, healthcare, telecommunications, transport or finance.
3. Do not break into computers or networks
- Do not attempt unauthorised access or get around security controls.
- Do not write or spread malware, steal credentials or launch denial-of-service attacks.
- Authorised security research is allowed, strictly within the scope of the permission you hold (see Part 3).
4. Do not develop weapons
Do not use the Service to design, make or obtain weapons, including chemical, biological, radiological and nuclear weapons and explosives.
5. Do not incite violence or hatred
Do not generate content that promotes violence, terrorism or extremism, or that incites hatred or discrimination on grounds such as ethnicity, race, religion, gender or disability.
6. Do not violate privacy or identity rights
- Do not collect, compile or disclose other people’s personal data without authorisation.
- Do not upload other people’s personal data or an organisation’s confidential information without authorisation.
- Do not use the Service to track or surveil a particular person, or to carry out facial recognition or identification without consent.
- Do not impersonate anyone.
7. Do not endanger minors
Do not generate any content that sexualises or sexually exploits minors, and do not use the Service to groom or harm a minor. Where we find such content, we may act on it as the law allows and report it to the competent authorities.
8. Do not harass, defraud or harm others
- Do not harass, threaten or intimidate anyone.
- Do not run scams or phishing, or forge identity documents, receipts or other documents.
- Do not generate intimate imagery of a person without their consent.
- Do not promote or encourage self-harm or suicide.
9. Do not create or spread disinformation
- Do not mass-produce and spread information you know to be false.
- Do not fake the source of information, manipulate public opinion or run networks of fake accounts.
- Do not interfere with elections or other public decision-making.
- Do not pass off AI-generated content as the work of a real person in order to deceive, and do not remove or alter AI-generated content labels.
10. Do not abuse the Service
- Do not get around usage, regional or security limits, evade content safety review, or try to recover masked private values or bypass approvals.
- Do not register accounts in bulk by automated means, or open accounts to harvest free allowances or invitation rewards.
- Do not share, resell or rent out accounts or allowances.
- Do not use cloud executors for anything unrelated to your task, such as cryptocurrency mining, running proxies or public-facing services, or launching network scans or attacks.
- Do not interfere with the Service or with other users.
- Do not use the Service to get around the terms of a third-party service, model, software or dataset.
- Do not block or evade client updates.
11. Do not generate sexually explicit content
Do not generate sexually explicit content.
12. Do not breach research ethics or academic integrity
- Do not fabricate or falsify research data or experimental results and present them as genuine.
- Do not invent citations and use them as if they were real references.
- For research involving human participants, animal experiments or controlled materials, obtain the ethics approvals and permits required.
Part 2. Requirements for high-risk uses
The uses below are not prohibited, but the results can affect someone’s health, rights, property or important opportunities. If you use the Service this way, you must meet the requirements in this part.
What counts as high-risk: Using outputs to make, or help make, a decision about a specific person in any of these areas:
- medical diagnosis and treatment;
- legal advice and legal proceedings;
- credit, insurance, investment and other financial decisions;
- hiring, performance review and dismissal;
- admissions, grading and academic assessment;
- eligibility for housing and public services.
Requirements:
- Review by a qualified person: Before an output is used or sent, someone with the relevant professional competence must review it and take responsibility for it.
- Tell the people affected: Let them know that AI was involved in the work.
- No full automation: Do not let the AI make a decision with a significant effect on a person without human review.
What does not count as high-risk: General learning and reference, literature reviews, internal drafts, and research analysis that is not about a specific person.
Part 3. Additional rules for specific situations
Agentic tasks and executors
- Commands, file operations and network requests the AI carries out for you in an executor count as your own actions.
- Before you start, confirm the execution location, working directory and approval mode. If you choose Full mode or automatic approval, you are responsible for what happens as a result.
- Do not let a task reach systems, data or accounts you are not entitled to access.
- For long-running or automatically executed tasks, you are responsible for overseeing what they do and what they cost.
- When running on your own device, grant the app only the system permissions the task needs.
Built-in browser and sign-in state
- Sign in to, or import, only accounts that you yourself are entitled to use. Do not import sign-in state from someone else’s browser.
- After an import, the AI can act as you on those websites. Import only the websites a task needs.
- Follow the terms of the websites you visit, including their limits on automated access and scraping.
- Do not use it for bulk registration, inflating metrics, automated purchasing, or getting around verification.
Private values (passwords, keys)
- Store and authorise only passwords and keys that you yourself are entitled to use.
- Do not hand another person’s credentials to a task unless you have been authorised to.
- Do not put passwords or keys in ordinary messages or in memory.
Voice input
Do not record other people’s conversations without their consent.
On-device models and your own model endpoints
- This policy applies equally to models that run on your phone and to models reached through an endpoint you enter yourself.
- When you use your own endpoint, you must also follow that provider’s terms.
Security research
- Get explicit authorisation from the owner of the target system first, and stay within it.
- Keep evidence of that authorisation and provide it if we ask.
- Do not use attack tooling produced by the research outside the authorised scope.
Providing content or services generated by the Service to others
- If you give outputs to other people, or interact with them by means of the Service, let them know the content is AI-generated or that they are interacting with AI.
- You are responsible for what you provide to others.
Minors
The Service is not designed for children. Age requirements are in section 2 of the Terms of Service. Do not use the Service to build products or services aimed at minors.
Part 4. How we enforce this policy
How we find problems: We run automated content safety review on inputs and outputs in the relevant features, and authorised staff may look at flagged material where needed. We also act on user reports and on requests from authorities. Section 3.11 of the Privacy Policy explains the personal data processing involved.
If your use does not comply with this policy, we may take any of the following steps without notice, as suits the situation:
- block a particular request or output;
- give a notice or warning;
- restrict some features or allowances;
- suspend the account and freeze its balance, allowances and any other funds paid into it;
- close the account, with no refund of purchased entitlements or paid-in funds except where the law requires otherwise;
- preserve evidence as the law requires and report to the competent authorities.
Notice and appeal: Where the law allows and it would not hinder a security response, we will tell you the reason and give you a way to contact us or appeal.
Part 5. Reporting and contact
If you come across content or conduct that breaks this policy, or you disagree with action we have taken, contact:
Include the task ID, when it happened and whatever explanation is needed. Redact sensitive information in screenshots and logs, and do not send passwords, verification codes, tokens or keys.
Part 6. Changes to this policy
We will update this policy as the law, the risks and the product change, and show the date. We will draw your attention to material changes through the product, our website or another reasonable means.