OpenAI Introduces Lockdown Mode to Boost Security

ISLAMABAD: (Web Desk) – OpenAI has unveiled a new security-focused feature called Lockdown Mode, designed to provide stronger protection against prompt injection attacks. These attacks occur when hidden or malicious instructions are embedded within web pages or other content sources, potentially influencing a chatbot’s behavior.

With Lockdown Mode enabled, ChatGPT’s access to live web content is significantly restricted. Users can only interact with cached web data, while image retrieval and display from online sources are disabled. The feature also limits certain advanced capabilities, including deep research tools and agent mode functions, to reduce potential security risks.

Iran Launches Ballistic Missile Attack on US Bases in Gulf

Despite these safeguards, OpenAI acknowledges that Lockdown Mode does not eliminate all threats. Prompt injections may still originate from cached content or user-uploaded files, which could affect response accuracy or system behavior.

According to the company, the primary purpose of Lockdown Mode is to minimize the possibility of sensitive information being unintentionally exposed through prompt injection techniques. OpenAI emphasized that the feature is intended for individuals and organizations handling confidential data who require stronger protections against data exfiltration risks.

The rollout of Lockdown Mode has begun for self-serve ChatGPT Business customers and select eligible personal account holders, with broader availability expected over time.