Skip to content
Submit incident
AII-1002-64517
AllegedThe record is a lawsuit, complaint or accusation that has not been decided.Supporting sources are pending verification.

DeepSeek's R1 model triggers market turmoil and intellectual property allegations

Chinese AI startup DeepSeek released advanced language models reportedly trained using model distillation with OpenAI's API outputs, allegedly violating OpenAI's terms. The release triggered significant US stock market losses and raised questions about intellectual property protection in AI development across jurisdictions.

January 2025 Β· DeepSeek Β· Intellectual property and appropriation
January 2025
AII-1002-64517
Quick facts
Category
Intellectual property and appropriation
Country
Singapore
Date
January 2025
Drafted by
Raidu Incident Detection

What happened

DeepSeek, a Chinese AI startup, released its R1 and V3 models, which reportedly achieved performance comparable to advanced Western models such as GPT-4 at significantly lower training costs. OpenAI and Microsoft alleged that DeepSeek used model distillationβ€”a technique where a smaller model learns from the outputs of a larger modelβ€”but specifically by using OpenAI's API outputs, which violates OpenAI's terms of service. The technique reportedly allowed DeepSeek to develop capable language models without the vast datasets and computational resources typically required.

Reported impact

The release of DeepSeek's models triggered approximately one trillion dollars in losses in the US stock market within a short timeframe. The development raised geopolitical concerns and challenged assumptions about American technological superiority in AI. The incident demonstrated that advanced AI model development may not require the extensive infrastructure and capital investment previously assumed necessary.

Index analysis

The allegation centers on intellectual property misuse through model distillation using OpenAI's proprietary API outputs without authorization. OpenAI claims this violates its terms of service and constitutes unauthorized use of its intellectual property. The case raises questions about whether model distillation using public APIs constitutes infringement and how intellectual property protections apply to AI model outputs in international contexts.

Classification

Category
Intellectual property and appropriation
Industry
Financial and insurance services
Country
Singapore
Harm status
Actual harm
Intent
Unknown
Timing
Unknown
Harm types
Economic or property
Harmed
Business
Failure modes
Training data misuse
AI systems
DeepSeek R1, DeepSeek V3, GPT-4

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident