Skip to content
Submit incident
AII-1000-40051
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

Employee Deletes 89TB of AI Data to Make Space for Unauthorized Work

A Chinese algorithm engineer intentionally deleted 89 terabytes of AI training data from his employer's servers to create storage space for unauthorized external work. After a 17-hour deletion process paralyzed the company's AI game development projects, he was convicted of destroying a computer information system and sentenced to five years and ten months imprisonment, ordered to repay over 200,000 yuan in damages.

3 September 2024 ยท Wang (็Ž‹ๆŸ) ยท Security and cyberattacks
3 September 2024
AII-1000-400512 sources โ†“
Quick facts
Category
Security and cyberattacks
Country
China
Source
ไบบๆฐ‘็ฝ‘
Date
3 September 2024
Drafted by
Raidu Incident Detection

What happened

On 3 September 2024, Wang, a 90s-born algorithm engineer, logged into a server cluster A using credentials from his previous role and entered a command to recursively delete all files and subdirectories without confirmation, commonly known as 'deleting database and running away' code.

The code ran for over 17 hours and deleted 89 terabytes of data from the AI game department, including source code, multiple generative text-to-3D and rendering AI models, and training data. The company discovered the incident on 4 September 2024, after which Wang was terminated.

Chat records between Wang and an external algorithm engineer revealed that Wang had been covertly using company computing resources to run model training for external work, and had deleted the company data to free up storage space for downloading his own private project models.

Reported impact

The deletion caused the company's model training system to stop functioning, and all AI game department research projects halted.

The company invested approximately 20 days of labour and computing resources costing over 40,000 yuan to recover the majority of deleted data, with total losses audited at over 160,000 yuan in employee wages plus recovery costs.

Index analysis

On 26 June 2025, Beijing's second-instance court upheld the original conviction that Wang's deletion was intentional, not accidental as Wang had claimed.

The court found that AI models, comprising hardware, software and data as an integrated system, constitute a 'computer information system' under criminal law, eligible for legal protection under the statute on destruction of computer information systems.

Governance controls

The company had issued three internal notices between 23 July and 5 August 2024 instructing all employees to migrate their data from the old cluster A to new cluster B, and notified that cluster A would no longer be accessible to the department after migration. Wang acknowledged receipt.

The Beijing Dongcheng District Procuratorate identified the core issue as clarifying Wang's subjective intent by examining his mobile phone data, discovering chat records showing awareness that using company resources without authorization would lead to dismissal.

Classification

Category
Security and cyberattacks
Industry
IT infrastructure and hosting
Country
China
Harm status
Actual harm
Intent
Intentional
Timing
Unknown
Harm types
Economic or property
Harmed
Business
Failure modes
Agent misaction
Consequences
Litigation, System suspension or termination, Content removal
AI systems
AI game training system, AI short-video models

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

ไบบๆฐ‘็ฝ‘
AIๆ—ถไปฃ๏ผŒ็ฎ—ๅŠ›่ต„ๆบๅฆ‚ไฝ•ไฟๆŠค
4 August 2026
ๆฏๆ—ฅ็ปๆตŽๆ–ฐ้—ป
ๅŒ—ไบฌ้ฆ–ไพ‹็ ดๅAIๆจกๅž‹ๅˆ‘ๆกˆ่ฝๆงŒ๏ผš"90ๅŽ"็ฎ—ๆณ•ๅทฅ็จ‹ๅธˆ"ๅˆ ๅบ“"่Žทๅˆ‘
6 August 2026