El Salvador Law for the Promotion of AI and Technologies (DL 234)

Binding law — in force

Tell users whether AI made or drove a decision, explain how it was reached, and let them challenge it before a competent person (El Salvador, DL 234 Art. 18)

DL 234, Art. 18 (AI decisions: notice, explanation and challenge before a natural person) · official text · In force: applies since 11 Mar 2025 · SV

El Salvador's Law for the Promotion of AI and Technologies (Legislative Decree 234) Art. 18 provides that where AI is used commercially or to access rights or services in the Republic, the user must be told whether the decision was taken directly by AI or driven by it (an AI-driven decision is one with a significant AI contribution, alone or with human operators, Art. 4(i)); the notification must contain understandable and transparent explanations of how the decision was reached, and there must be mechanisms to challenge it before a competent natural person who can confirm, modify or revoke it. Compliance is a condition for the Law's safeguards (Art. 19). Detect a model-driven decision sent to the person with no AI-decision notice or challenge route.

Who it applies to

• Duty falls on: any person
• Systems covered: automated decision, consequential decision
• Anyone who uses AI commercially, or to decide access to rights or services, in El Salvador: every decision taken directly by AI or driven by it (a significant AI contribution, Art. 4(i)) is notified with that fact and an understandable explanation, and can be challenged before a competent natural person who can confirm, modify or revoke it. In force since 2025-03-11. The Law has no sanctions chapter: compliance conditions the Art. 19 safeguards. Whether other law makes Art. 18 enforceable, and what 'commercially' covers, are questions for counsel (review flag).
• Whether it applies depends on facts outside the code; a person has to decide.

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.')

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}'))

Standards that recommend the same control

• Significant AI decisions should be documented and contestable with remedies (CoE Framework Convention) (CoE AI Framework Convention (CETS 225) · CoE Framework Convention on AI (CETS 225), Article 14(2))
• Financial institutions should disclose AI use and explain AI-driven decisions on request (MAS FEAT) (MAS FEAT Principles · MAS FEAT Principles — Transparency, Principles 12-14)
• Health AI/LMMs should be transparent and documented before deployment (WHO) (WHO AI-for-Health (LMM) Guidance · WHO guidance on large multi-modal models (2024), Executive summary (Figure 1), section 1.2 and Box 1: the six consensus principles of WHO's 2021 guidance)

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

Rule id sv-dl-234.ai-decision-notice-explanation-and-human-challenge · review status: primary source derived