A standard for accountable AI

Why HAL exists

HAL emerged from a simple observation. Human-in-the-Loop was designed for assistants. Human Accountable for the Loop was designed for systems. As AI evolves from generating outputs to taking actions, reviewing every decision becomes impractical, so the question has to change.

Instead of asking "was a human involved?", HAL asks "who is accountable for the system that made the decision?"

Where it fits

HAL does not compete with the standards organisations already follow. It complements them. NIST helps you understand AI risk. ISO 42001 helps you build an AI management system. The EU AI Act defines regulatory obligations. HAL answers a different question: when can an organisation responsibly delegate decisions and actions to AI while retaining clear human accountability?

That question becomes more important with every step from copilots to agents. HAL exists to answer it.

What we believe

• Human-in-the-Loop is not the same as accountability.
• Accountability can never be delegated. Execution can.
• The future of governance is owning the system that makes decisions.

Author

HAL is a framework by Ryan McDonough. It is published openly: read the theory, explore the framework, and put your systems to the test with the assessment and calculator.