Binding law — in force AI-adjacent law
Inform people of solely automated decisions and, on request, hear their view and have a natural person review the decision (Switzerland, FADP Art. 21)
Under Switzerland's Federal Act on Data Protection (FADP/DSG, SR 235.1) Art. 21, a controller must inform the data subject of any decision based exclusively on automated processing that has a legal consequence for them or significantly affects them (an automated individual decision); on request it must let them state their point of view, and they may require the decision to be reviewed by a natural person. This does not apply where the decision is directly connected with concluding or performing a contract and the request is granted, or where the person expressly consented to the automated decision. A private person who wilfully fails to inform is liable to a fine of up to CHF 250,000 (Art. 60). Detect a model-driven adverse decision with no automated-decision notice or no human-review route.
Trust and provenance not reviewed by a lawyer · audit-grade · source verified 3 Oct 2026 · release 2026.10.03.3
- Lane
- Binding law — in force In force: applies since 1 Sep 2023
- Official source
- DSG Art. 21(1)-(3) (duty to inform of an automated individual decision; view and human review; exceptions) · captured 2 Oct 2026 · anchor hash (SHA-256)
4e3500ecc7e2…· 5 more anchors in the data release - Verification
- Quoted text found word for word in the captured official document (3 Oct 2026). Source last verified 3 Oct 2026: checked against the captured official document; not in the weekly watcher's list; checked against the captured document.
- Data release
- Data release 2026.10.03.3, data as of 3 Oct 2026, schema 0.3.9.
- Legal review
- Not reviewed by a lawyer. TwinEthos derived this rule from the official text it cites: treat it as research to check against that text; it is not legal advice. No TwinEthos rule has been legally reviewed yet. Open questions for counsel on this rule: 1.
- Audit standard
- Audit-grade: meets all 10 checks of the TwinEthos audit standard that apply to it. The audit standard is TwinEthos's own quality bar for provenance, dates, applicability, detectors, fixtures, remediation and licences; it is not a legal review.
- Detectors
1 detector (code pattern), experimental: written from the rule's text and not yet measured for precision on real code, so treat a hit as a lead to verify.
Known limits:
- Adverse outcomes as numeric codes
- Decision set in a different file from the model call
- The notice and review route may live in a letters or portal module; follow the decision path before reporting.
Who it applies to
- Duty falls on: controller
- Systems covered: automated decision
- Private controllers (and federal bodies) whose decisions based exclusively on automated processing of personal data have a legal consequence for, or significantly affect, people, where the facts have effects in Switzerland (Art. 3(1)): inform them of the automated individual decision; on request let them state their view and have a natural person review it; contract (request granted) and express-consent exceptions. In force since 2023-09-01. The English text is not official; the German (or French or Italian) text governs. What 'erheblich beeinträchtigt' covers and whether an AI-assisted decision with nominal human sign-off is 'ausschliesslich automatisiert' are questions for counsel (review flag).
- Not covered:
- A decision directly connected with concluding or performing a contract with the data subject where their request is granted (Art. 21(3)(a))
- The data subject expressly consented to the decision being automated (Art. 21(3)(b))
- Personal data processed by a natural person exclusively for personal use (Art. 2(2)(a))
- Processing by the Federal Assembly and parliamentary committees, and by institutional beneficiaries with immunity (Art. 2(2)(b)-(c), read, not stored)
- Whether it applies depends on facts outside the code; a person has to decide.
The guard to add
Route significant automated decisions through meaningful human review, or wire in an automated-decision notice, reasons, human intervention, a way to give a view, and contest.
At the point where model output becomes a significant decision about a person (approve, deny, underwrite, set_status), either queue the case for a reviewer who weighs the evidence and can change the outcome before it takes effect (review_queue.enqueue, requires_human_review), or, where the decision stays solely automated, record the permitted basis for that decision type and wire the safeguards in. Those safeguards are a notice in the decision message that it was made by automated processing, reasons the person can read, and request_human_review or contest routes where the person can give their view and have a human reconsider. A reviewer who approves every case without examining it does not make the decision non-automated, so the review records reviewer identity, the evidence viewed, and the outcome.
Where it goes: 1 application source code, 9 AI output handling, 15 agent action surface, 14 user-facing text.
What this provision adds:
- Inform the person that the decision is an automated individual decision when it is communicated.
- On request, let the person state their point of view and have a natural person review the decision.
Example (Python + OpenAI SDK), before:
verdict = client.chat.completions.create(model=MODEL, messages=msgs).choices[0].message.content
if verdict.strip() == 'deny':
deny(applicant)
send_decision_email(applicant, 'Your application was not approved.')After:
out = client.chat.completions.create(model=MODEL, messages=msgs,
response_format={'type': 'json_object'})
result = json.loads(out.choices[0].message.content)
if result['decision'] == 'deny':
if requires_human_review('credit'): # a person decides
review_queue.enqueue(applicant.id, proposal=result)
else: # solely automated, recorded basis
deny(applicant, basis=DECISION_BASIS['credit'], reasons=result['reasons'])
send_decision_email(applicant, render('adm_denial.txt', notice=ADM_NOTICE,
reasons=result['reasons'], contest_url=f'{BASE}/decisions/{applicant.id}/contest'))Control: Solely-automated significant decision without human-intervention safeguards. The same guard addresses 15 items with binding law in 15 jurisdictions. Engineering guidance, not legal advice.
Related incidents
No guardrail sits on this exact control; these incidents are cited by guardrails on related controls.
- UnitedHealth nH Predict claim-denial litigation (2023-11; alleged (not proven)). A class action filed in November 2023 alleges that UnitedHealth's nH Predict model had a 90% error rate, measured by denials reversed on appeal, while only about 0.2% of members appealed. UnitedHealth disputes the allegations; the litigation is ongoing. Source: STAT News · evidence grade: primary · cited by Monitor how often adverse AI decisions are reversed, and suspend models that are usually wrong
- Cigna PXDX batch claim denials (reported) (2022; alleged (not proven)). ProPublica, citing internal Cigna records, reported that Cigna's PXDX system was used to reject more than 300,000 claims over two months in 2022, with physicians spending an average of 1.2 seconds on each. Cigna disputes the reporting; related lawsuits are ongoing. Source: ProPublica / The Capitol Forum · evidence grade: press of record · cited by Make human review of adverse AI decisions substantive, not nominal
Rule id ch-fadp.automated-individual-decision-information-and-review · review status: primary source derived