OpenAI has introduced a new feature called Lockdown Mode, designed to bolster protection against prompt injection attacks, which involve embedding harmful instructions within web content. This mode disables live web browsing, restricting users to cached content, and prevents the retrieval of images from the internet, although users can still generate images. Despite these measures, OpenAI cautions that Lockdown Mode does not eliminate the risk of prompt injections entirely, as vulnerabilities could still arise from cached content or uploaded files. The primary objective of Lockdown Mode is to minimize the risk of sensitive data being inadvertently shared. OpenAI emphasizes that this feature is not universally applicable; it is specifically intended for individuals and organizations that manage sensitive data and require enhanced security measures against data exfiltration risks. Currently, Lockdown Mode is being rolled out to self-serve ChatGPT Business accounts and selected personal accounts, marking a significant step in OpenAI's commitment to data security.