DeepSeek AI suffered a major security breach, exposing over a million log entries, secret keys, and sensitive database details. The leaked information could have allowed unauthorized access to its internal systems.
A security report revealed that DeepSeek left its ClickHouse database open online. This database permitted full control over its operations, allowing attackers to access confidential data without authentication. The leaked logs contained chat histories, backend details, API secrets, and other operational metadata.
The exposed database, hosted on DeepSeek’s subdomains, enabled direct execution of SQL queries through a web browser. This flaw allowed complete database control and potential privilege escalation within DeepSeek’s environment. However, it remains unclear if malicious actors accessed or downloaded the data before the issue was fixed.
A researcher emphasized that AI services must prioritize security as rapid adoption increases risks. While futuristic AI threats often make headlines, real dangers come from basic security lapses, such as unprotected databases. Companies must ensure that AI engineers and security teams collaborate to prevent such incidents.
DeepSeek has gained massive popularity for its open-source AI models, which rival leading AI systems while being efficient and cost-effective. Its chatbot surged to the top of app store rankings but soon became a target of large-scale attacks, forcing it to pause new registrations.
The company is also under scrutiny for its privacy policies and ties to China, raising national security concerns in the U.S. Additionally, Italian regulators blocked its apps, questioning its data handling practices. Meanwhile, OpenAI and Microsoft are investigating whether DeepSeek used OpenAI’s API without permission to train its models through a technique called distillation.
Preventive Measures
To prevent such breaches, AI companies must enforce strict data security measures. They should regularly audit access controls, encrypt sensitive data, and restrict database exposure. Collaborating with cybersecurity teams is essential to prevent unauthorized access and safeguard user information.