T7

Representation & Equity

Where people reach government through AI agents, a private layer sits between what a user meant to say and what the agency receives. That layer changes the submission: it may arrive better ordered, or stripped of the details that made it recognizably that person’s. Either effect depends on the agent, which the service neither supplies nor can inspect.

The users with the least room to adjust are those with no agent, a weak one, or one built on a model that represents them poorly. They are often the users a service most needs to reach. An agency that keeps handling submissions as they arrive lets the quality of a user’s agent decide whose case it hears well.

01

Policy challenge

Government owes people roughly equal standing to reach an outcome it controls, whether to be heard in a consultation or to transact for a benefit, license, payment, or appeal. That equality has rested on users reaching the service directly.

More participation now runs through a privately supplied agent layer that government neither distributes nor controls. Whether a user's request lands depends on three things the service itself does not set: whether they have a capable agent at all, how well a model represents people like them, and whose interests the agent serves in advancing their position.

If an agency kept handling requests as they arrive, standing in its own services would track the quality of a user's tooling. The agency could not then meet its duty to treat the public on equal terms without acting on the agent layer directly.

02

Design challenge

Make sure every user can reach a government outcome, whether to be heard in a consultation or to transact for what they are entitled to, whether they bring a capable agent or none at all.

Keep what the user said intact regardless of the intermediary.

Make any agent route usable whatever a user's language, literacy, or assistive-technology needs.

Guarantee a route without an agent that reaches the same outcome, not just the same access.

Account for who a model serves badly, and keep those users capably represented.

Patterns in this territory

9 shown
7.1 Frontier

Agency-provided agents (a public option)

A free, government-provided agent at every service entry point, aligned to the user and auditable by the public. A user without a bank, employer, or subscription that supplies an agent still gets one at the door of the service.

7.2 Frontier

Preserving the person's own voice

Keeping what the user said intact through an agent's rewriting: the original travels with the polished version, and meaning-changing edits are flagged before anything is sent. What the person said remains on the record, available whenever the polished version is disputed.

7.3 Emerging

Non-agent fallback channels, kept first-class

A non-agent route to every outcome, held at genuine parity of quality and timeliness with the agent route. A person who declines the agent route still gets the benefit, license, or decision on the same terms and timetable.

7.4 Frontier

Accessibility baseline: WCAG and beyond

Holding the agent interface itself to the recognized accessibility floor, and extending that floor to how the agent converses, not only how it renders.

7.5 Emerging

Plain language and multilingual intake

Agent intake that works across literacy levels and languages: plain language by default, an honest indicator of how well the agent handles the user's language, and a human interpreter when it doesn't. That disclosure stops the service from settling consequential matters in a language the agent only appears to understand.

7.6 Emerging

Co-design with affected communities

Giving the communities a service is built for a hand in how its agent talks, before it ships and after: co-designed conversational behavior, and standing oversight of the live system by the people it affects. Failure modes only an affected community would recognize get raised while they can still be fixed.

7.7 Frontier

Whose interests the platform serves

Making visible whose interests a commercially supplied agent serves when it acts for a user, with an uncompromised public alternative one action away. A user deciding whether to rely on a supplied agent can see whose interests it serves besides their own.

7.8 Frontier

Government-provided agents for high-stakes interactions

A capable, government-provided agent for high-stakes interactions (an appeal, a dispute, an enforcement action), offered to users who lack their own, with a human escalation path throughout. The outcome of an appeal or enforcement action stops tracking whether the person could afford a capable agent.

7.9 Emerging

Digital divide and exclusion risk

Naming, before an agent channel launches, the populations it will fail, and committing outreach and a working alternative route for each. A population the channel will fail is provided for before launch instead of surfacing later in complaint and harm data.

Case studies that touch this territory