SGC

Safety Guidance Compliance — Official policy and framework for implementing, monitoring, and enforcing safety measures, risk mitigations, and responsible AI usage across all Blue Helm Studios platforms, models, and services.
1. Purpose

The Safety Guidance Compliance (SGC) policy establishes mandatory standards and procedures to identify, mitigate, and manage safety risks associated with the use of Blue Helm Studios AI models and platforms. It ensures responsible development, deployment, and usage while protecting users, society, and the integrity of our systems.

2. Scope

This policy applies to all Blue Helm Studios AI models, APIs, applications, developer tools, and integrated services. It covers internal teams, external users, business partners, and any third parties interacting with our technology.

3. Core Safety Principles

All usage of Blue Helm Studios technology must align with these principles:

  • Do No Harm – Prevent foreseeable risks of physical, psychological, or societal harm
  • Respect Human Autonomy – Avoid manipulation, deception, or undue influence
  • Transparency – Clearly disclose when content or interactions involve AI
  • Fairness & Non-Discrimination – Mitigate biased or harmful outputs
  • Accountability – Maintain audit trails and responsibility for safety measures
4. Safety Requirements

Blue Helm Studios and its users/partners must implement:

  • Robust safety filters and refusal mechanisms for high-risk prompts
  • Red-teaming and adversarial testing procedures
  • Real-time monitoring for emerging safety risks
  • Clear user guidance on safe and responsible prompting
  • Incident reporting and post-mortem analysis for safety events
5. Prohibited High-Risk Uses

The following uses are strictly prohibited under SGC:

  • Development or deployment of weapons, explosives, or biological agents
  • Autonomous decision-making in critical infrastructure (healthcare, transportation, energy, etc.) without human oversight
  • Manipulation of elections, public discourse, or large-scale disinformation campaigns
  • Creation of deepfakes intended to cause harm or deceive the public
  • Any use that could lead to catastrophic outcomes or existential risks
6. Enforcement and Compliance

Non-compliance with safety guidance may result in immediate suspension or termination of access. Blue Helm Studios reserves the right to audit usage, request safety documentation, and report serious violations to relevant authorities. Partners must maintain records of their safety measures for review.

7. Reporting and Support

Safety incidents, near-misses, or questions about responsible use should be reported to safety@bluehelmstudios.com. We encourage proactive feedback to continuously improve our safety frameworks.

Document ID: BHSLS-SGC-001 • Version 1.0 • Effective Date: April 8, 2026