Declared automation on outbound decisions
The outbound decision itself declares what was automated and to what degree, in a form both the person and their agent can read. Reliance can then be calibrated decision by decision.
The impact of agents
As agents come to issue more of government's outbound decisions, the person holding one will be less and less able to tell, from the artifact itself, whether and how far automation produced it. The disclosure duties now arriving are entity-level: a privacy policy states the kinds of decisions a computer program makes somewhere in the organization, a fact the person will never connect to the letter in their hand. Reliance then miscalibrates in both directions at once: machine output is trusted as if an officer weighed it, or every notice is distrusted because none says who decided.
The first reader of an outbound decision may be the person's own agent, deciding whether to flag, contest, or act on it. A declaration buried in prose serves that reader as poorly as boilerplate serves a human one; without a per-decision, machine-checkable declaration, the receiving agent must guess at the very fact it needs to calibrate.
What must be verified
A person, and the agent reading on their behalf, needs to be able to tell from the decision artifact itself whether automation made the decision, materially shaped it, or merely assisted. The artifact should say which of these, using the same automation vocabulary the service publishes elsewhere. That per-decision declaration then has to be checkable against the agency's entity-level disclosures. The agency issuing the decision holds the duty to place the declaration on every outbound decision, to keep it accurate as the automation behind the decision class changes, and to keep it consistent with what its privacy policy and transparency statements claim.
Protecting access
When a declaration uses technical or legal vocabulary, it excludes readers with low literacy or numeracy, limited English, or cognitive load, at exactly the moment an adverse decision arrives. For them the mention becomes one more line of boilerplate, and their reliance stays miscalibrated at equal effort. A machine-readable declaration on its own tilts the other way, serving only whoever already has an agent to parse it.
Keeping the path open
- Make the human-readable line primary and concrete: 'this assessment was produced automatically; no officer reviewed it before it was sent'.
- Render that line identically across every channel the decision travels, including phone scripts and paper.
- Comprehension-test it with low-literacy, limited-English, and cognitive-accessibility cohorts.
- Treat the machine-readable duplicate as an addition for the person's agent, never a substitute for the line a person can read.
- Make sure screen readers reach the declaration wherever it renders.
Response surface
Each decision carries its own declaration of what was automated and whether an officer reviewed it, rather than pointing at the policy behind it.
Rebate assessment
DN-2026-1187Your rebate for this quarter is $118.40.
This assessment was produced automatically. No officer reviewed it before it was sent.
The declaration line renders identically on every channel the decision travels — letter, phone script, and portal. It states only how far automation acted; the reasons for the decision and the path to contest it are separate duties the declaration feeds.
Maturity
- Established
For the per-decision mention, in force and judicially enforced in France since 2016 under CRPA article L311-3-1, treated by the Conseil constitutionnel in 2020 as an obligation other provisions must satisfy.
- Emerging
For the dated statutory duties either side of it: Australia's entity-level APP 1.7-1.9 disclosure commencing 10 December 2026, and EU AI Act Article 26(11), deferred to 2 December 2027.
- Frontier Headline
As a calibration instrument, stating the level of automation behind each decision in the service's own graded vocabulary, with a machine-readable duplicate for the person's agent, which no jurisdiction yet operates.
Precedents
France, CRPA article L311-3-1, the mention on the individual decision. The article requires an individual decision taken on the basis of algorithmic processing to include an explicit mention informing the person, with the processing rules and the principal characteristics of its implementation communicated on request. The Conseil constitutionnel treats the obligation as live law that an admissions-secrecy provision had to satisfy. Decision-level declaration is operating practice in one jurisdiction.
Australia, APP 1.7-1.9, the entity-level baseline. The Privacy and Other Legislation Amendment Act 2024 requires a privacy policy to disclose the kinds of personal information used by computer programs that make decisions, the kinds of decisions made solely by them, and the kinds substantially assisted. A policy-level disclosure tells the population that automation exists. It cannot tell the person that automation produced the decision they are holding.
EU AI Act Article 26(11), deployer notice to the person. Deployers of Annex III high-risk systems that make or assist decisions about natural persons must inform those persons that they are subject to the system. It is adopted law with a commencement date, and operating practice nowhere.
Canada's Directive on Automated Decision-Making, notice through every channel. The Directive requires plain-language notice that a decision will be undertaken wholly or in part by an automated system, at every impact level, through all service delivery channels in use. Channel parity is written into the duty rather than left to implementation.
What carries over to agent use
The French mention proves the per-decision declaration is administrable at national scale: the line has been mandatory on every covered decision since 2016, and the constitutional court has enforced the obligation's presence rather than treating it as decoration. Canada's channel rule and Australia's solely-versus-substantially distinction both transfer directly as content for the declaration: together they supply the where (every channel, every covered decision) and the what (which grade of automation acted).
What breaks in transit is the reader. Every existing instrument addresses a human reader; none contemplates the person's own agent as the first consumer of the declaration, so the machine-readable duplicate, its format, and its binding to the human-readable line are undesigned. The declaration also does only its own job: it states that and how far automation acted. The reasons behind the decision, any path to contest it, and whether it arrives before or after the decision takes effect are a separate duty this declaration feeds into, not something it provides itself.
Where things go wrong
Undeclared, people rely on machine output as if a person had weighed their circumstances, and the absence of review is invisible until an error surfaces downstream. The declaration can be forged in the flattering direction: a decision stamped as officer-reviewed on the strength of a seconds-long open-and-close, borrowing the label's credibility for output no one weighed. It can also decay while every line stays honest: the per-decision declaration and the entity-level disclosures drift out of sync as the automation behind a decision class changes, until the person's agent reads one story on the decision and another in the policy. Keeping the declaration accurate, and the two layers consistent, is the duty that closes both gaps.
Sources
5 references
The instrument, the operating deployment, or the official record itself.
- Légifrance — Code des relations entre le public et l'administration, article L311-3-1
- Conseil constitutionnel — Decision no. 2020-834 QPC (Parcoursup algorithms)
- Privacy and Other Legislation Amendment Act 2024 (Cth), as passed — APP 1.7-1.9
- Treasury Board of Canada — Directive on Automated Decision-Making (notice rows, all impact levels)