Our Policy on Frontier Safety (Inflection)

• Authority: Private-sector companies
• Collections: Corporate policies and commitments
• Most recent activity: Enacted
• Most recent activity date: 2023-10-30

Risk subdomain coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the risk subdomain are explicitly described.

• AI system security vulnerabilities
• Cyberattacks, weapon development or use
• Fraud, scams, and targeted manipulation
• Competitive dynamics
• Governance failure
• AI possessing dangerous capabilities
• Lack of capability or robustness

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the risk subdomain are described.

• Exposure to toxic content
• Compromise of privacy
• False or misleading information
• Disinformation, surveillance, and influence at scale
• Overreliance and unsafe use
• AI pursuing its own goals
• Lack of transparency or interpretability

All other risk subdomains had no coverage in this document.

Risk Coverage Summary

The document has good coverage of approximately 12-14 subdomains, with strong focus on AI system security (2.2), malicious actors (4.1, 4.2, 4.3), AI safety failures (7.1, 7.2, 7.3), governance (6.5), competitive dynamics (6.4), and misinformation (3.1). Coverage is concentrated in security, misuse prevention, AI safety, and governance domains.

Sector coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the sector are explicitly described.

• Information
• Scientific Research & Development Services

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the sector are described.

• Health Care & Social Assistance
• National Security

All other sectors had no coverage in this document.

Sectors Summary

This is an internal corporate policy document from Inflection, an AI development company. The primary sector governed is Information (specifically AI development and data processing services) and Scientific Research and Development Services, as these are the sectors in which Inflection operates. The document also addresses child safety concerns, touching on aspects relevant to Educational Services and Health Care sectors through discussion of mental health applications.

Legislative Status

Other

Legislative Status Summary

This is an internal corporate policy document outlining Inflection's voluntary commitments and safety practices for frontier AI development. It uses primarily voluntary language ('believes', 'intends', 'supports') and describes self-imposed standards without external enforcement mechanisms.

Actors

Proposer

• Entity Types: AI_Developer
• Entities: Inflection
• Summary: This is Inflection's own policy document describing their internal safety practices and commitments. The document is authored by and represents the commitments of Inflection as an AI development company.

Target

• Entity Types: AI_Developer
• Entities: Inflection
• Summary: The policy applies to Inflection's own operations, model development, and deployment practices. It is an internal governance framework for the company's AI development activities.

Enforcer

• Entity Types: AI_Developer
• Entities: Inflection safety team; Inflection legal team; Chief Executive Officer
• Summary: Enforcement is conducted internally by Inflection's own safety team, legal team, and CEO. The safety team has authority to block model launches and implement mitigations.

Monitor

• Entity Types: AI_Developer
• Entities: Inflection safety team; external security researchers (bug bounty program); domain expert red-teamers
• Summary: Monitoring is conducted by Inflection's internal safety team with support from external security researchers and domain experts through red-teaming and bug bounty programs.

AI Lifecycle Stages

Lifecycle stages covered by the documents, as defined in the OECD Framework for the Classification of AI Systems

• Plan and Design
• Collect and Process Data
• Build and Use Model
• Verify and Validate
• Deploy
• Operate and Monitor

AI Lifecycle Summary

The document comprehensively covers all stages of the AI lifecycle with particular emphasis on pre-deployment evaluation (Plan and Design, Verify and Validate), deployment processes, and post-deployment monitoring (Operate and Monitor). It addresses data collection practices, model building, testing, deployment, and ongoing monitoring with detailed governance measures at each stage.

Technical Scope

Terms relating to technical scope that are mentioned or implicitly referenced in the document

• AI Models
• Frontier AI
• AI Systems
• Generative AI

Technical Scope Summary

The document explicitly focuses on frontier AI models and systems, with repeated references to 'frontier AI' throughout. It does not explicitly define or distinguish between general purpose AI, task-specific AI, or foundation models, though the context suggests focus on large-scale language models. No specific compute thresholds are mentioned, and there is no explicit discussion of open-weight models.