The Race — Building Democracy in the Age of AI. People race toward a shared finish line.
The Race
← Tools

Practical tools for contestable scale

Accountability and due process

What makes AI actions visible, reviewable, and correctable?

Registers, notice and appeal, independent testing, incident reporting, action logs, and delegated permissions.

13 policy tools. Status distinguishes verified use from adopted measures, proposals and ideas.

13 of 13 policy tools match · Showing 12

Policy tool · See record

Agent action logs and delegated-authority controls

Bucket: Accountability and Due-Process Tools / Anti-Lock-In Tools Status: Public tool / active tool candidate Primary lanes: Lane 05 — Platforms, Distribution, and Separations; Lane 01 — Governance and Regulation

Current status: Illustrative instrument; current implementation and access limits qualified

Lanes 1, 3, 5, 6 · Source snapshot: 2026-11-20

Read tool and examples →

Policy tool · See record

Trusted-user transparency standards for frontier AI access

Bucket: Accountability and Due-Process Tools / Public Capacity Tools Status: Emerging / operationally grounded Primary lane: Lane 02 — National Security and Frontier Control

Current status: Illustrative instrument; current implementation and access limits qualified

Lanes 1, 3, 5, 6 · Source snapshot: 2026-10-07

Read tool and examples →

Policy tool · See record

Public AI registries

Give the public a searchable account of government AI uses so people can discover which systems affect them and identify the responsible institution. A registry supports scrutiny and requests for explanation; it is not itself an audit or an appeal process.

Current status: In use

Lanes 1, 2, 5, 6 · Source snapshot: 2026-10-31

Read tool and examples →

Policy tool · See record

Notice and appeal rights for consequential AI decisions

Tell people when a consequential decision is automated and give them a meaningful route to challenge it. Depending on the law, this may include an explanation, human intervention, a chance to present their position, a regulator complaint or judicial review.

Current status: Enacted rights; coverage varies

Lanes 1, 2, 5 · Source snapshot: 2026-10-25

Read tool and examples →

Policy tool · See record

Independent testing and evaluation

Have people outside the supplier’s development team test an AI system’s performance and risks. Give evaluators enough access, clear methods and the ability to report adverse findings so purchasing and oversight do not depend entirely on vendor assurances.

Current status: Operating METR access with protected publication; independence varies by arrangement

Lanes 1, 2, 5, 6 · Source snapshot: 2026-10-17

Read tool and examples →

Policy tool · See record

AI incident reporting

Require developers or operators to report specified serious AI incidents to a responsible authority within defined deadlines. Reporting can reveal failures that would otherwise stay private and give regulators a basis for investigation and corrective action.

Current status: Enacted requirements; implementation varies

Lanes 1, 2, 5, 6 · Source snapshot: 2026-10-22

Read tool and examples →

Policy tool · England and Wales

Institutional AI approval and escalation gates

Require a named public institutional authority to approve tools before consequential deployment and provide a route for escalation, while retaining human responsibility.

Current status: Adopted institutional policy in England and Wales, October 9; prior human-checking guidance exists. No new litigant appeal right or evaluated effectiveness established.

Lanes 1, 6, 7 · Source snapshot: 2026-10-13 · Linked evidence assessment

Read tool and examples →

Policy tool · International

Publisher AI content usage reporting

Use a common specification to report when AI agents access or use publisher content so institutions can inspect usage and negotiate terms.

Current status: SPUR v1.0 specification available September 2 under Apache 2.0; reporting relies on agent cooperation. Audited completeness, remuneration and smaller-publisher outcomes are not established.

Lanes 5, 1, 7 · Source snapshot: 2026-10-13

Read tool and examples →

Policy tool · United States

Citizen suits to enforce environmental requirements

Where legislation authorizes citizen enforcement, affected people and organizations can ask a court to enforce environmental obligations even when executive officials prefer not to proceed. Standing, notice, statutory limits and remedies determine actual access; the instrument applies to AI through its physical infrastructure.

Current status: See current status

Lanes 1, 3, 6, 2 · Source snapshot: 2026-10-21

Read tool and examples →
Related Specific Events (5)