All legal documents

Responsible AI Use

Last updated: March 16, 2026

1. Our Commitment

sh0 is committed to developing and deploying AI responsibly. We believe AI should be beneficial, safe, and fair. This document outlines our ethical AI principles and our expectations for how users should use sh0's AI features.

2. AI Ethics Principles

2.1 Beneficence

We design sh0's AI features to create positive value for users:

  • Help developers deploy faster and more reliably
  • Reduce time spent debugging build and deployment errors
  • Optimize resource usage to reduce costs
  • Democratize access to DevOps expertise

2.2 Fairness & Non-Discrimination

We strive to make sh0's AI fair to all users:

  • No discrimination based on protected characteristics
  • Equal service quality for all users regardless of plan
  • Addressing bias in AI outputs
  • Testing for disparate impact across technology stacks

2.3 Transparency

We are transparent about AI capabilities and limitations:

  • Clear disclosure of AI usage in every feature
  • Honest about what AI can and cannot do
  • AI suggestions always labeled as such
  • Confidence indicators where applicable

2.4 Accountability

We take responsibility for our AI systems:

  • Clear responsibility for AI outputs
  • Responding to user concerns promptly
  • Fixing discovered problems
  • Learning from mistakes and improving

2.5 Privacy & Security

We protect user data and privacy:

  • Code sent to AI is not stored or used for training
  • Self-hosted users control all AI interactions
  • AI features can be completely disabled
  • Compliance with privacy laws

3. Responsible Use Guidelines

3.1 For Deployment

sh0's AI is designed to assist with deployment tasks:

  • Use AI diagnostics to understand build errors
  • Apply AI-suggested fixes after review
  • Use resource recommendations as a starting point
  • Leverage Chat-to-Deploy for rapid prototyping
  • Always test AI-generated configurations before production

3.2 For Decision-Making

Consider AI as a tool to assist, not replace human judgment:

  • Use AI-generated suggestions as a starting point
  • Apply your expertise and judgment
  • Verify important recommendations
  • Test changes in staging before production
  • Take responsibility for final decisions

3.3 For Production Systems

For production deployments:

  • Do not blindly apply AI-suggested changes to production
  • Always review configuration changes before deployment
  • Maintain rollback procedures
  • Monitor systems after applying AI recommendations
  • Use AI analysis as supplementary input, not the sole basis

4. Prohibited Uses

sh0's AI features must not be used for:

4.1 Malicious Purposes

  • Generating malicious code or exploits
  • Identifying vulnerabilities for exploitation
  • Creating tools for unauthorized access
  • Circumventing security measures

4.2 Abuse

  • Overwhelming AI services to degrade performance for others
  • Extracting or reverse-engineering AI model weights
  • Reselling AI outputs as a standalone service
  • Using AI to generate spam or abusive content

5. Bias & Fairness

We acknowledge that AI can reflect biases:

5.1 Bias Types

  • Training Data Bias: AI learns from historical data with existing biases
  • Stack Bias: AI may be more accurate for popular technology stacks
  • Language Bias: AI may perform differently across programming languages

5.2 Our Mitigation

  • Regular evaluation across diverse technology stacks
  • Prompt engineering to reduce bias
  • User feedback on accuracy and fairness
  • Continuous improvement based on real-world usage

5.3 User Responsibility

  • Review AI outputs for accuracy
  • Report biased or inaccurate suggestions
  • Apply professional judgment
  • Do not assume AI suggestions are always correct

6. Environmental Impact

We are mindful of AI's environmental footprint:

  • Efficient Models: Using cost-effective models (DeepSeek V3) that balance quality and resource usage
  • On-Demand Processing: AI is only invoked when requested, not continuously
  • Non-AI Alternatives: Many features (Code Health Check, stack detection) use pure Rust without AI
  • Green Infrastructure: Cloud providers committed to renewable energy

7. Accessibility

We are committed to making sh0 accessible:

  • WCAG 2.1 AA compliance for the dashboard
  • Screen reader support
  • Keyboard navigation
  • Clear, plain-language AI explanations

8. Reporting Concerns

If you encounter problematic AI behavior:

  • Email: [email protected]
  • Report bias, harmful suggestions, or inaccurate outputs
  • Include: description, context, impact
  • We will investigate and respond within 7 days

9. Ongoing Commitment

sh0 commits to:

  • Regular ethics reviews of AI features
  • Staying current with AI safety research
  • Engaging with the developer community on AI ethics
  • Listening to user feedback
  • Continuously improving AI responsibility
  • Publishing transparency information in release notes

10. Contact & Questions

Questions about responsible AI?