Coupang Allegedly Tweaked Search Algorithms to Boost Own Products

Coupang was alleged in internal reports tampering its search algorithms to prioritize exposure of its own products, which potentially violated Korea's Fair Trade Act.

Risk classification

• Primary risk domain: 6 Socioeconomic & Environmental
• Primary risk subdomain: 6.1 Power centralization and unfair distribution of benefits

Coupang used its market dominance and search algorithm control to prioritize its own products, centralizing economic power and unfairly distributing benefits to itself over third-party suppliers.

Additional risk subdomains

• 4.3 Fraud, scams, and targeted manipulation: The companies manipulated search results to deceive consumers and gain an unfair financial advantage over competitors.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The algorithmic manipulation was a deliberate decision made by humans at Coupang and Naver to prioritize their own products post-deployment.

EU AI Act risk tier

• Risk tier: 4 Minimal or No Risk

Minimal or No Risk: The search and recommendation algorithms used by Coupang and Naver do not fall under the high-risk or limited-risk categories defined by the EU AI Act, representing applications with minimal regulatory obligations under this framework.

AI system and alleged parties

• AI system: Coupang search algorithm
• AI purpose: Content Search; Content Recommendation
• Behaviour type: Tool
• Alleged developer: Coupang
• Alleged deployer: Coupang
• Alleged harmed parties: Coupang suppliers, Coupang customers

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Substantial. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Substantial, indirect Substantial
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Minor, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report explicitly describes significant financial losses, including a 3.29 billion won ($2.81 million) fine for Coupang, a 26.7 billion won ($23 million) fine for Naver, and 5.7 billion won ($4.8 million) in promotional expenses forced onto 388 suppliers.

Directly caused: The average direct financial loss per occurrence includes the regulatory fines and direct supplier burdens, totaling approximately $15.3 million per occurrence across the two documented cases.

Indirectly caused: Third-party suppliers suffered indirect financial losses due to reduced exposure and lost sales from being pushed to the bottom of search results.

Inferred additional harm: It is highly likely that hundreds of other third-party vendors suffered substantial unquantified financial losses due to decreased search visibility and unfair competition.

Differential treatment

Reported: The report explicitly describes differential treatment where Coupang's search algorithm prioritized its own private-label products over those of third-party suppliers.

Directly caused: Coupang's private-label products were placed at the top of search results, while products from other companies were pushed to the bottom, receiving less exposure.

Inferred additional harm: Systematic disadvantage was likely caused to thousands of small and medium-sized merchants who rely on fair algorithmic ranking to compete on these platforms.

Epistemic

Reported: The report explicitly describes epistemic harm, noting that Naver 'deceived consumers' and Coupang 'tampered with algorithms' to alter search results.

Directly caused: Consumers were presented with manipulated search results that did not reflect organic relevance, misleading them about product popularity and pricing.

Inferred additional harm: Widespread erosion of consumer trust in the integrity of search results and online marketplaces is highly likely.

People affected

• Occurrences reported: 2
• People reportedly harmed: 388
• People reportedly exposed: 388

Potential causes

Management

• Prioritizing Profit Over Fairness: Management directed algorithmic bias to improve margins on private labels.

Technology

• Algorithm Manipulation: Search algorithms were altered to prioritize private label products.
• Search Result Tampering: Algorithmic tweaks pushed competitor products to the bottom of pages.

Data Inputs

• Biased Search Ranking Inputs: Input parameters favored private label metadata over merchant performance.

Human Factors

• Intentional Algorithm Meddling: Developers manually tweaked search parameters to favor internal products.

Process and Methods

• Unfair Search Optimization Methods: Methods used for ranking lacked fair-trade compliance checks.

Regulatory Environment

• Inadequate Algorithmic Auditing: Lack of regulatory oversight allowed algorithm tampering to go undetected.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide consistent and clear details regarding the regulatory investigations, the specific algorithmic manipulation (prioritizing private labels), and the resulting fines. The role of the AI system (the search algorithm) is explicitly stated, and the financial impacts on both the companies and the suppliers are well-documented.
• Ambiguities identified: The exact technical mechanism of how the search algorithm was tweaked is not detailed, and the reports conflate the algorithmic manipulation with other non-AI business practices like forcing suppliers to buy ads.
• Alternative interpretations: The incident could be viewed purely as a traditional antitrust/fair trade violation rather than an AI safety failure, though the central role of algorithm manipulation makes it highly relevant to AI governance.

South Korean e-commerce giant Coupang manipulated its search algorithms to prioritize private-label products over third-party competitors. While the incident represents a significant domestic antitrust issue resulting in regulatory fines, it has negligible national security implications and does not threaten critical infrastructure, state sovereignty, or national economic security.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: South Korea
• Alleged perpetrator: Coupang

Threat characteristics

• Imminence: Long-term. The incident represents a resolved regulatory matter with no active or immediate threat to national security.
• Autonomy: Human-controlled. The algorithmic ranking manipulation was directly programmed and controlled by human operators to achieve specific corporate objectives.
• Novelty: Established threat. Self-preferencing and algorithmic manipulation by dominant e-commerce platforms are well-established antitrust issues globally.

Impact by dimension

• Physical security: Negligible. No physical systems, infrastructure, or human safety were impacted by the commercial search algorithm manipulation.
• Information security: Negligible. The incident involves commercial search manipulation rather than state-sponsored disinformation, intelligence compromise, or information warfare.
• Sovereignty: Negligible. No threat to state authority or government functions; the incident was successfully managed through South Korea's standard regulatory and legal frameworks.
• Economic security: Minor. The incident caused financial harm to local suppliers and distorted market competition, but it represents a domestic fair trade issue rather than a strategic threat to national economic security.
• Societal stability: Negligible. No significant impact on societal stability, civil liberties, or human rights, as the harm was limited to commercial unfairness and consumer deception.