Control
Adverse AI decision without explanation/appeal
An adverse consequential AI decision must come with reasons, an appeal path, and (where feasible) data correction.
Informational data, not legal advice. Summaries are TwinEthos's own words and rules have not been reviewed by a lawyer: check the official text before relying on any of it. A guard addresses an item; adding it is not a statement that your code meets any law.
Reach
enacted, not yet applying in European Union (EU), Colorado (US-CO); next date 2027-01-01.
The guard to add
Send each adverse AI-assisted decision with its main reasons and the AI's role, plus a way to correct data and appeal to a human who can change the outcome.
Where model output becomes an adverse status (denied, declined, rejected, ineligible), the decision service stores reason codes or principal reasons, the model id and version, and an input snapshot or hash with the decision. The notice to the person (letter, email, portal response) says AI was involved and what role it played, lists the main factors, and links to data correction and to an appeal that creates a human-review task with authority to change the outcome. An explanation endpoint returns the stored record on request, so the deployer can explain a decision long after the model has changed.
Where it goes: 2 data models, 9 AI output handling, 14 user-facing text.
What reviewers look for: in files that set an adverse status from model output, reason_codes, principal_reasons, key_factors, or role_of_ai stored with the decision; an adverse-action template or response that carries those reasons, the AI's role, a data-correction link, and an appeal route; an appeal endpoint that creates a human_review task.
Example (Python + OpenAI SDK + Pydantic), before:
resp = client.chat.completions.create(model=MODEL, messages=msgs)
if 'deny' in resp.choices[0].message.content.lower():
application.status = 'denied'
send_email(applicant.email, 'Your application was declined.')After:
a = Assessment.model_validate_json(resp.choices[0].message.content) # decision, reason_codes
if a.decision == 'deny':
decisions.insert(app_id=application.id, status='denied', reason_codes=a.reason_codes,
model=resp.model, input_hash=hashlib.sha256(payload).hexdigest())
send_email(applicant.email, render('adverse_action_notice.txt',
reasons=a.reason_codes,
role_of_ai='An AI model assessed your application; a reviewer can change the outcome.',
correct_data_url='/profile/data', appeal_url=f'/appeals/new?decision={application.id}'))Engineering guidance, not legal advice. Each provision below may add its own details (a cadence, a deadline, a required notice element): open it for those.
Upcoming dates
- : Adverse ADMT outcomes require a 30-day plain-language explanation (Colorado (US-CO); first application; stayed)
Every rule this guard addresses
Binding law — not yet in force or stayed (2)
- Colorado (US-CO)
- Adverse ADMT outcomes require a 30-day plain-language explanation C.R.S. 6-1-1704(3) · applies from 2027-01-01 · stayed
- European Union (EU)
- Affected persons can demand a meaningful explanation of a high-risk AI decision (EU AI Act Art. 86) Article 86 · applies from 2026-08-02
Standard / soft law (3)
- Everywhere (*)
- Health AI/LMMs should be transparent and documented before deployment (WHO) WHO Ethics and Governance of AI for Health (LMMs, 2024) — Six Consensus Principles
- International (INTL)
- Significant AI decisions should be documented and contestable with remedies (CoE Framework Convention) CoE Framework Convention on AI (CETS 225), Article 14(2)
- Singapore (SG)
- Financial institutions should disclose AI use and explain AI-driven decisions on request (MAS FEAT) MAS FEAT Principles — Transparency, Principles 12-14
TwinEthos recommendation (not law) (1)
- Everywhere (*)
- Explain adverse AI-assisted decisions and offer a way to contest them — everywhere TwinEthos derivation — guardrail.baseline-adverse-decision-explanation
Related incidents
- 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 Explain adverse AI-assisted decisions and offer a way to contest them — everywhere