# The Institute for Ethical AI Alignment & Safety > An independent research institute working to ensure that frontier AI is safe, aligned and open for people and society. ## Overview - [Talks & Keynotes on Production ML, MLOps and AI Security](https://ethical.institute/talks-and-events/): Keynotes, conference talks and roundtables on production ML, security, explainability and agentic systems. - [Contact the Institute](https://ethical.institute/contact/): A single route into the network, frameworks, research and institutional partnerships. ## Policy & Frameworks - [AI Policy & Standards — EU AI Act, UK, UN and ISO Engagement](https://ethical.institute/policy/): Practitioner evidence, public-policy products and standards participation across European, UK, UN and US processes. - [The Nine Principles for AI Alignment & Safety](https://ethical.institute/principles/): The nine principles for AI alignment and safety: commitments, failure modes and the practical controls that implement them, for teams deploying, integrating and fine-tuning AI systems. - [AI-RFX — Open Procurement Framework and RFP Templates for AI](https://ethical.institute/frameworks/ai-rfx/): Open-source templates that convert the Nine Principles for AI Alignment & Safety into a procurement checklist for organisations buying, building or deploying AI systems. - [The State of Production ML 2025 — MLOps Practitioner Survey](https://ethical.institute/reports/state-of-ml-2025/): Practitioner-reported production machine learning practices, tools, organisational context and demographics from the 2025 survey. - [AI Safety Frameworks — Procurement, Maturity, Security](https://ethical.institute/frameworks/): Procurement, maturity and security frameworks that translate the Nine Principles for AI Alignment & Safety into practice. ## Open source - [KAOS — Kubernetes Agent Orchestration System for AI Agents](https://ethical.institute/open-source/kaos/): Deploy, manage and orchestrate AI agents on Kubernetes as declarative resources, secured at the gateway and observable end to end with KAOS. - [Kompute — Vulkan GPU Compute Framework (Linux Foundation)](https://ethical.institute/open-source/kompute/): A fast, mobile-enabled and asynchronous general-purpose GPU compute framework for cross-vendor graphics cards from AMD, Qualcomm, NVIDIA and others. - [XAI — Explainability and Bias Toolbox for Machine Learning](https://ethical.institute/open-source/xai/): XAI is a machine learning library for explainability across data analysis, model evaluation and production monitoring, bringing tools and process together. - [Production Machine Learning — 545 Open-Source MLOps Tools](https://ethical.institute/open-source/production-ml-list/): A curated catalogue of 545 open-source libraries across 24 categories for deploying, monitoring, versioning and scaling machine learning. - [Open-Source AI and MLOps Tools from the Institute](https://ethical.institute/open-source/): Openly licensed software for agent orchestration, GPU compute, explainability and production ML. - [AI Regulation & Guidelines — Global Policy Map](https://ethical.institute/open-source/ai-guidelines/): An open catalogue mapping AI guidelines, principles, ethics codes, standards and regulation, from national law to practical checklists. - [Agent Skills Marketplace for Claude Code & Copilot CLI](https://ethical.institute/open-source/agent-skills-marketplace/): An open-source marketplace of portable agent skills and plugins built on the SKILL.md convention, installable in both Claude Code and Copilot CLI. - [Production Agentic Systems: Open-Source Agent Tools](https://ethical.institute/open-source/production-agentic-list/): A curated catalogue of 79 open-source libraries across 7 categories for deploying, monitoring, securing and scaling production agentic systems. - [Monokai Institute — Neovim and Vim Colorscheme](https://ethical.institute/open-source/monokai-institute/): A truecolor Monokai colorscheme for Neovim and Vim, with a light sage variant, transparency support and highlights tuned for LSP and Treesitter. ## Initiatives - [Machine Learning Maturity Model — Eight Assessment Criteria](https://ethical.institute/frameworks/maturity-model/): Make safe and aligned ML repeatable through a practical path from isolated good intentions to adaptive organisational capability. - [MLSecOps — Machine Learning and Agentic AI Security Guidance](https://ethical.institute/frameworks/security/): MLSecOps security guidance for the machine-learning lifecycle, extended to agentic systems with delegated authority, tools and memory. - [Agentic RFX — Procurement Criteria for Agentic AI Systems](https://ethical.institute/frameworks/agentic-rfx/): Procurement criteria for agentic systems, in development as an extension of the AI-RFX framework. - [Agentic Maturity Model — Capability Levels for AI Agents](https://ethical.institute/frameworks/agentic-maturity-model/): Organisational capability levels for operating agentic systems, currently in development. - [The State of Production ML 2024 — MLOps Practitioner Survey](https://ethical.institute/reports/state-of-ml-2024/): Practitioner-reported production machine learning practices, tools, organisational context and demographics from the 2024 survey. - [State of Production ML Surveys — Methodology and Citation](https://ethical.institute/reports/): The 2024 and 2025 State of Production ML practitioner surveys, with respondent counts, field dates, methodology and citation guidance. - [Partners — Institutions Working with the Institute](https://ethical.institute/partners/): Organisations work with the Institute through frameworks, open source, the network, and standards and policy engagement. ## Network - [Ethical AI Network — Practitioners in ML, Policy & Standards](https://ethical.institute/network/): The ML Engineer newsletter and Ethical AI Network connect practitioners working across engineering, research, policy and standards. - [Ethical AI Network membership](https://ethical.institute/membership/): Membership in the Ethical AI Network is free and built through contribution to shared principles, frameworks and open-source work. - [The ML Engineer newsletter](https://ethical.institute/newsletter/): 406 issues, 2018-12-16 to 2026-09-27. - [Production ML, responsible AI and alignment blog](https://ethical.institute/blog/): 17 long-form articles, 2020-09-12 to 2026-09-27. ## About - [About the Institute for Ethical AI Alignment & Safety](https://ethical.institute/about/): Learn about the Institute's history, mission and independent work to make frontier AI safe, aligned and open for people and society. - [Branding: the Institute logo, mark, colour and type](https://ethical.institute/about/branding/): The identity of The Institute for Ethical AI Alignment & Safety: wordmark and mark lockups to download, the brand palette, and the typeface behind them.