IMDA Safety Testing Starter Kit v1.0 Aligned
Acceptable Use Policy (AUP)
Effective Date: July 23, 2026 | Version 2.0 | Legal Inquiries: legal@letschat.asia
🛡️ Policy Summary & Core Restrictions
- Prohibited AI Conduct: Prompt injection attacks, jailbreaking, deceptive identity impersonation, and unauthorized web harvesting are strictly prohibited.
- Safety Red Teaming: Let's Chat evaluates deployments against IMDA's Starter Kit for Testing LLM-Based Applications for Safety and Reliability (v1.0).
- Enforcement: Violations result in immediate token revocation and account suspension.
- Reporting Contact: legal@letschat.asia.
1. Purpose & Prohibited Conduct
This Acceptable Use Policy specifies permitted and restricted practices for deploying chatbots and AI agents on Let's Chat. Users and tenants must not execute or facilitate:
- Adversarial Attacks: Submitting prompt injections or jailbreak vectors designed to bypass system guardrails.
- Deceptive Impersonation: Falsely representing AI agents as specific living human individuals without authorization.
- Resource Exhaustion: Infinite agent loop executions or unapproved high-frequency scraping.
- Illegal & Harmful Content: Ingesting or generating defamation, hate speech, or intellectual property infringements.
2. AI Safety Testing Standards
Let's Chat incorporates automated security evaluations aligned with IMDA's safety testing guidelines:
Prompt Injection Defense
Automated filtering of jailbreak attempts and instructions altering core system prompts.
Groundedness & Accuracy
RAG similarity scoring prevents ungrounded hallucinations from being delivered to end-users.
3. Enforcement & SLA
Accounts violating this AUP are subject to immediate token invalidation. Security incidents or misuse reports should be submitted to:
Email: legal@letschat.asia (24-hour response SLA for critical safety reports)