Acceptable Use Policy
Last updated: 25 June 2026
This policy applies to everyone using Harbourstone. Because our AI features run on Anthropic’s Claude API, your use must also comply with Anthropic’s Usage Policy; the obligations below pass through to you.
One rule above all: don’t use Harbourstone — or the Claude features within it — to do anything illegal, harmful, deceptive, or anything Anthropic’s Usage Policy prohibits.
1. Prohibited uses
You must not use the service or its AI features to:
- Generate or distribute malware, ransomware, denial-of-service or other malicious code, or tools to bypass security controls or intercept communications.
- Bypass, jailbreak or circumvent safety guardrails, restrictions or capabilities of Claude or our products.
- Produce fraudulent, falsified or deceptive content, or impersonate others.
- Collect, expose or misuse personal or private information without permission and a lawful basis.
- Generate content that is illegal, that exploits or endangers minors, that promotes violence, harassment or hate, or that Anthropic’s Usage Policy otherwise prohibits.
- Scrape, reverse-engineer, distil or train competing models on the underlying model, or resell raw model access.
- Infringe intellectual-property or other rights.
2. High-risk decisions require a human
Harbourstone is for reporting and visualisation. You must not use AI outputs to make, or to substantially automate, decisions that materially affect people’s rights or access to services — in particular:
- Finance: investment advice, loan approvals, creditworthiness or financial-eligibility decisions.
- Employment: resume screening, hiring tools, or other employment determinations.
- Insurance and similar: coverage or eligibility decisions affecting individuals.
If your use ever approaches these areas, a qualified person must review outputs (human-in-the-loop) and you must disclose that AI was involved in producing them.
3. Geography
AI features may not be used by, or made available to, users located in regions excluded by Anthropic’s terms (including Russia). Harbourstone is intended for Kazakhstan and Central Asia.
4. Enforcement
We may suspend or terminate access for violations, and may report unlawful activity. Violations may also breach Anthropic’s terms and result in loss of AI features.
Questions or reports of misuse: abuse@harbourstone.kz.