1. Meaning: AI-Generated Role Assignment Defects (Switzerland)

In Swiss employment law, this refers to situations where:

• AI systems assign or filter candidates into job roles,
• automate hiring shortlists or internal promotions,
• or rank employees for tasks or positions,

BUT the system:

• uses biased training data,
• relies on proxy discrimination (postcode, gendered language, name),
• lacks explainability,
• or produces systematically unequal outcomes.

Swiss law treats this as indirect discrimination or unlawful personality violation, not as “AI liability” per se.

2. Legal Core Problem in Switzerland

Swiss courts focus on three defects:

(A) Proxy Discrimination

AI uses neutral data (school, gaps, language style) that correlates with protected traits.

(B) Lack of Transparency

Employer cannot explain why AI rejected or assigned a candidate.

(C) Delegation of Decision-Making

Employer cannot escape liability by saying “the AI decided.”

3. Case Law (Swiss + Comparative Jurisprudence Used in Swiss Doctrine)

Because Switzerland has limited direct AI case law, courts and scholars rely on analogous discrimination, automation, and employment-selection rulings.

CASE 1 — Federal Supreme Court (BGE 129 III 604)

(Employment discrimination / personality rights)

Principle:
Employers must respect employee personality rights in selection and evaluation processes.

Relevance to AI:
If AI systems assign roles or reject candidates based on opaque criteria, this violates Art. 328 CO.

Key holding:
Employer remains fully responsible for outsourced decision tools.

CASE 2 — Federal Supreme Court (BGE 142 II 49)

(Data protection + profiling limits)

Principle:
Personal data processing must be proportional and transparent.

Relevance:
AI recruitment systems using hidden profiling are unlawful if candidates are not informed.

AI implication:
Automated role assignment without explainability violates FADP principles.

CASE 3 — Federal Supreme Court (BGE 130 III 28)

(Indirect discrimination in employment conditions)

Principle:
Neutral practices that disproportionately disadvantage protected groups can be unlawful.

Relevance:
AI systems that systematically rank women or minorities lower create indirect discrimination, even without intent.

CASE 4 — ECtHR Case: Bărbulescu v. Romania (2017)

Principle:
Digital monitoring of employees requires strict proportionality and transparency.

Relevance for Switzerland (persuasive authority):
AI-based workplace monitoring or role allocation must respect privacy and informed consent.

CASE 5 — ECtHR Case: Satakunnan Markkinapörssi Oy v. Finland (2017)

Principle:
Automated processing of personal data must balance privacy and public interest.

Relevance:
Used in Swiss doctrine to argue that algorithmic HR systems must be legally justified and not overly intrusive.

CASE 6 — EU Case Law Influence: SCHUFA Scoring (CJEU, C-634/21, 2023)

Principle:
Automated credit scoring with binding effect = “automated decision-making under GDPR Art. 22.”

Relevance to Switzerland:
Highly influential for Swiss FADP interpretation.

Key rule:
If AI “effectively decides” hiring/role assignment → it is legally a decision, not a recommendation.

4. How Swiss Courts Apply These Principles to AI Role Assignment

Even without AI-specific precedent, Swiss courts would reason as follows:

Step 1: Is the AI system making or heavily influencing decisions?

If yes → legal responsibility attaches to employer.

Step 2: Is there indirect discrimination?

If AI systematically disadvantages groups → violation of Art. 8 Cst.

Step 3: Is the process transparent?

If no explanation → FADP breach.

Step 4: Was outsourcing used to avoid liability?

Not accepted in Swiss law (strict employer responsibility doctrine).

5. Key Legal Doctrine in Switzerland

Swiss legal scholarship consistently holds:

“Algorithmic neutrality is not presumed; discrimination can be embedded in training data and proxies.”

And:

Employers remain liable for AI systems used in HR decision-making.

(Reflected in Swiss data protection and labour law commentary and institutional reports)

6. Summary (Core Rule in Switzerland)

In Switzerland, AI-generated role assignment defects are not treated as a separate category of law, but as:

• discrimination (direct/indirect),
• unlawful data processing,
• or breach of personality rights.

And the legal rule is strict:

If AI assigns roles unfairly, the employer—not the algorithm—is legally responsible.