Library Borrowing Recommendation Liability in Singapore

(Algorithmic Recommendations, User Harm Claims, Data Responsibility & Institutional Duty)

“Library borrowing recommendation liability” refers to legal issues that arise when libraries (public or academic) or their digital systems:

• recommend books or materials via algorithms or librarians
• influence borrowing choices
• collect and process borrowing history
• potentially cause harm through inaccurate, biased, or inappropriate recommendations

In Singapore, this area is not governed by a specific “library recommendation law.” Instead, liability is assessed through a combination of:

• administrative law principles (for public libraries)
• negligence and duty of care concepts
• data protection obligations (PDPA)
• contractual and institutional policies
• evidential fairness in disputes

1. What “Liability” Issues Can Arise?

(A) Harmful or misleading recommendations

• recommending outdated or inappropriate material
• algorithmic bias in reading suggestions
• misinformation amplification

(B) Privacy breaches via borrowing history

• exposing sensitive reading habits
• profiling users based on borrowing patterns

(C) Algorithmic discrimination

• unequal recommendation quality across user groups
• cultural or language bias

(D) Systemic reliance harm

• users relying on library recommendation systems for academic or professional decisions

(E) Institutional negligence claims

• failure to maintain safe, accurate recommendation systems
• lack of safeguards for minors or vulnerable users

2. Legal Nature of Library Recommendation Systems in Singapore

Libraries in Singapore (especially public institutions) are typically treated as:

• statutory/public service bodies
• subject to administrative law standards
• and indirectly subject to PDPA obligations for personal data handling

Recommendations themselves are generally considered:

• informational services, not binding decisions
• but may still attract liability if negligence or unfairness is proven

3. Core Legal Principles Applied

Courts typically evaluate such claims under:

• Wednesbury unreasonableness (irrationality)
• procedural fairness
• duty of care (negligence framework)
• confidentiality obligations
• legality of data usage
• judicial deference to institutional expertise

4. Relevant Case Laws (At Least 6)

Below are key Singapore and persuasive authorities relevant to recommendation systems, institutional liability, and data-driven decision risks.

1. Ridgewood Properties Pte Ltd v Land Management Corporation of Singapore [2005] SGCA

Principle:

• Courts will not substitute their judgment for administrative or technical decisions unless irrational or unlawful.

Relevance:

Library recommendation systems (especially algorithmic ones) are treated as:

• technical institutional tools
• courts defer heavily unless recommendations are clearly irrational

2. Chee Siok Chin v Attorney-General [2006] SGCA

Principle:

• High threshold for judicial review under Wednesbury unreasonableness.
• Courts avoid interfering with administrative discretion.

Relevance:

To establish liability for bad recommendations, claimant must show:

• extreme irrationality
  not just disagreement with suggested materials

3. Lines International Holding (S) Pte Ltd v Singapore Tourist Promotion Board [1997] SGHC

Principle:

• Authorities must follow fair, consistent procedures.
• Decisions must align with stated policies.

Relevance:

If library systems:

• deviate from stated recommendation criteria
• apply inconsistent categorisation

this may constitute procedural unfairness or breach of legitimate expectation.

4. Chng Suan Tze v Minister for Home Affairs [1988] SGCA

Principle:

• Executive discretion is reviewable for legality and rationality.
• Courts can scrutinise whether discretion is properly exercised.

Relevance:

If recommendation systems effectively determine access to materials:

• decisions must still be lawful
• cannot be arbitrary or unchecked

5. Tay Choon Seng v Public Service Commission [1993] SGHC

Principle:

• Administrative decisions must consider relevant factors and avoid arbitrariness.

Relevance:

Library recommendation systems may be challenged if:

• they ignore user context (age, academic need, language preference)
• or rely on irrelevant data signals

6. Borissik Svetlana v Urban Redevelopment Authority [2009] SGCA

Principle:

• Courts defer to specialised institutional judgment in technical matters.

Relevance:

Library professionals and algorithmic systems are considered:

• specialised knowledge-based decision-makers
• courts rarely second-guess recommendation logic

7. Council of Civil Service Unions v Minister for the Civil Service (GCHQ case) [1985] UKHL (Persuasive Authority)

Principle:

• Judicial review grounds: illegality, irrationality, procedural impropriety.
• Recognises confidentiality in sensitive systems.

Relevance:

Used to balance:

• confidentiality of recommendation algorithms
  vs
• fairness and accountability obligations

8. R v Panel on Take-overs and Mergers, ex parte Datafin plc [1987] UKCA (Persuasive Authority)

Principle:

• Private bodies performing public functions can be subject to judicial review.

Relevance:

If library recommendation systems are outsourced:

• private vendors may still be held accountable
• if they perform public-facing informational functions

5. Key Liability Conflict Areas

(A) Algorithmic bias vs neutrality duty

Libraries are expected to be neutral knowledge providers. Conflicts arise when:

• recommendation engines systematically favour certain genres, languages, or ideologies

Under Lines International, inconsistency may raise fairness concerns.

(B) Privacy of borrowing history vs personalised recommendations

To improve recommendations, systems often use:

• reading history
• search queries
• borrowing patterns

This conflicts with:

• confidentiality expectations
• PDPA data minimisation principles

(C) Harm from reliance on recommendations

If users rely heavily on library suggestions and suffer harm:

• liability may be argued under negligence principles
  But courts are cautious because:
• recommendations are generally informational, not authoritative advice

(D) Automated recommendation vs human oversight

A major issue is whether:

• algorithm alone generates suggestions
  or
• librarians review or adjust outputs

Lack of oversight may increase risk of unfair or harmful recommendations.

(E) Exposure of sensitive reading patterns

Borrowing history may reveal:

• medical interests
• political views
• religious materials

Unauthorized exposure may trigger:

• confidentiality breach concerns
• PDPA enforcement issues

6. Practical Legal Position in Singapore

Courts generally:

• treat library recommendations as informational, not binding decisions
• defer strongly to institutional expertise
• require high threshold for establishing liability

However, liability may arise where:

• there is clear negligence in system design or data handling
• confidentiality of borrowing history is breached
• recommendation systems operate arbitrarily or without safeguards
• statutory or PDPA obligations are violated

7. Conclusion

Library borrowing recommendation liability in Singapore is governed not by a single legal regime but by an intersection of:

• administrative law fairness principles
• negligence and duty of care concepts
• data protection obligations under PDPA
• judicial deference to institutional expertise

While courts are highly reluctant to treat recommendations as legally binding decisions, liability can still arise where:

• systems are biased or irrational
• user data is mishandled
• or institutional safeguards are inadequate

Overall, Singapore law strikes a balance between innovation in digital library services and accountability for fairness, privacy, and rational decision-making.