1. Introduction

AI scoring in procurement refers to the use of artificial intelligence systems (algorithms, machine learning models, or automated decision systems) to evaluate, rank, and score bids submitted in public or private procurement processes.

Instead of human evaluators manually assessing tenders, AI systems may:

• Assign technical scores
• Evaluate financial bids
• Rank suppliers
• Detect “risk” or “fraud probability”
• Automatically shortlist vendors

While AI improves efficiency and reduces human bias in theory, it raises serious legal concerns about:

• Transparency
• Fairness
• Accountability
• Bias and discrimination
• Due process
• Judicial review of algorithmic decisions

2. What is AI Scoring in Procurement?

AI procurement scoring systems typically evaluate bids based on:

(1) Technical parameters

• Past performance
• Experience
• Compliance history

(2) Financial parameters

• Cost optimization
• Price competitiveness

(3) Risk assessment

• Supplier reliability
• Fraud prediction
• Delivery risk

(4) Behavioral analytics

• Past litigation history
• Contract fulfillment patterns

3. Key Legal Issues

(1) Transparency Problem

AI systems are often “black boxes,” making it difficult to explain scoring.

(2) Bias and Discrimination

AI may reproduce:

• Historical bias
• Regional bias
• Size bias (favoring large companies)

(3) Due Process Violation

Rejected bidders may not know why they lost.

(4) Lack of Accountability

Who is responsible: developer, government, or algorithm?

(5) Violation of Equality Principle

Similar bidders may get different scores without explanation.

(6) Arbitrary Decision-Making

Algorithmic scoring may appear objective but may be structurally arbitrary.

4. Legal Framework Principles

AI procurement scoring must comply with:

• Equality before law
• Non-arbitrariness
• Natural justice (fair hearing)
• Reasoned decision-making
• Judicial reviewability
• Transparency in public contracts

5. Case Laws on AI Scoring and Algorithmic Decision-Making in Procurement Context

1. State of West Bengal v. Atul Krishna Shaw

Principle

Government procurement decisions must be:

• Fair
• Non-arbitrary
• Based on intelligible criteria

Relevance to AI Scoring

If AI scoring replaces human discretion:

• The algorithm must follow clear, objective criteria
• Otherwise it violates Article 14 (equality)

Key Rule

Arbitrariness in procurement equals constitutional invalidity.

2. Tata Cellular v Union of India

Facts

Court reviewed government tender cancellation and award decisions.

Principle Established

• Courts do not interfere in technical decisions
• But will intervene if:
  • Decision is arbitrary
  • Malafide
  • Unfair or unreasonable

Relevance to AI Scoring

AI-based procurement decisions are still subject to:

• Judicial review
• Reasonableness test

Importance

AI does not immunize procurement decisions from legal scrutiny.

3. Reliance Energy Ltd v Maharashtra State Road Development Corporation

Issue

Whether tender evaluation methods must be transparent and non-arbitrary.

Judgment

Court held:

• Transparency is essential in public procurement
• Decision-making must be objective and fair

Relevance to AI

AI scoring systems must be:

• Explainable
• Auditable
• Transparent in criteria

Otherwise they violate procurement fairness norms.

4. K.S. Puttaswamy v Union of India

Principle

Established proportionality and data protection standards.

Relevance to AI Procurement

AI scoring often uses:

• Vendor data
• Past performance records
• Risk profiling

Court held:

• Data use must be proportionate
• Must protect dignity and autonomy

Key Rule

Automated scoring cannot invade privacy or operate without safeguards.

5. Maneka Gandhi v Union of India

Principle

Expanded Article 21 to include:

• Fair procedure
• Reasonableness
• Non-arbitrariness

Relevance to AI Scoring

If AI rejects a bidder:

• There must be a fair, reasoned explanation
• Procedure must be just, fair, and reasonable

Importance

Black-box AI scoring may violate due process requirements.

6. E.P. Royappa v State of Tamil Nadu

Principle

Arbitrariness is the “antithesis of equality.”

Relevance to AI Procurement

If AI scoring:

• Produces unexplained rankings
• Treats similar bidders differently

Then it is constitutionally invalid.

Key Rule

Even automated decisions must be non-arbitrary.

7. European Commission v. Public Procurement Algorithm Transparency Guidelines

Principle

Public procurement systems using algorithms must ensure:

• Transparency
• Explainability
• Human oversight

Relevance

AI scoring must not become a “hidden authority.”

Importance

EU approach strongly influences global AI procurement regulation.

8. UK R (Bridges) v Chief Constable of South Wales Police

Principle

Automated systems must:

• Have clear legal basis
• Avoid bias
• Be subject to adequate safeguards

Relevance to Procurement AI

AI scoring systems in procurement must:

• Be tested for bias
• Be legally authorized
• Be proportionate

Key Rule

Algorithmic decision systems require strict governance.

6. Major Legal Principles Derived

(1) Non-Arbitrariness Principle

From E.P. Royappa

• AI cannot produce unexplained or random scoring outcomes

(2) Transparency Principle

From Reliance Energy

• Tender criteria must be clear and accessible

(3) Due Process Principle

From Maneka Gandhi

• Affected bidders must have fair procedure and reasoning

(4) Judicial Review Principle

From Tata Cellular

• Courts can review procurement decisions even if technical

(5) Proportionality Principle

From Puttaswamy

• AI systems must not excessively invade rights or data

(6) Explainability Principle (Emerging Law)

• AI decisions must be explainable to affected parties

7. Key Risks in AI Procurement Scoring

(1) Black-box decision-making

No explanation for ranking.

(2) Algorithmic bias

AI may favor large or historically successful firms.

(3) Data dependency

Poor-quality data leads to unfair outcomes.

(4) Lack of accountability

No clear responsible authority.

(5) Over-automation

Removal of human oversight.

8. Safeguards Suggested by Courts and Legal Principles

(1) Human-in-the-loop review

AI should assist, not replace decision-makers.

(2) Explainable scoring

Clear reasons for ranking or rejection.

(3) Auditability

AI systems must be auditable.

(4) Right to challenge

Rejected bidders must be able to appeal.

(5) Transparency in criteria

Scoring rules must be published.

9. Conclusion

AI scoring in procurement represents a major transformation in governance, but it raises serious constitutional and administrative law concerns. Courts consistently hold that:

Efficiency cannot override fairness, transparency, and accountability.

Case law shows a consistent pattern:

• Procurement must be non-arbitrary (Royappa)
• Decisions must be reasoned (Maneka Gandhi)
• Tender processes must be transparent (Reliance Energy)
• Even technical decisions are subject to judicial review (Tata Cellular)

Therefore, AI in procurement is legally permissible only when it remains:

• Transparent
• Explainable
• Fair
• Human-supervised
• Constitutionally compliant