OpenAI Extends ChatGPT Lockdown Feature to Free Accounts
openai
| Source: Mastodon | Original article
OpenAI expands ChatGPT Lockdown Mode to free users. The feature blocks outbound network requests to prevent data theft.
OpenAI has expanded its ChatGPT Lockdown Mode to free users, a move aimed at blocking prompt injection attacks that can lead to data theft. This feature, initially available only to enterprise users, restricts outbound network requests, preventing hackers from stealing sensitive information.
The expansion of Lockdown Mode to all account types, including Free, Plus, and Pro, underscores the growing concern over prompt injection attacks. These attacks involve hiding malicious instructions on external webpages that can hijack the language model, allowing hackers to secretly steal data. By making Lockdown Mode widely available, OpenAI is providing users with an additional layer of security against these advanced threats.
As the use of AI models like ChatGPT becomes more widespread, the risk of such attacks is likely to increase. OpenAI's decision to roll out Lockdown Mode to all users is a significant step in enhancing the security of its platform. Users can enable Lockdown Mode from their settings, adding a crucial defense against prompt injection attacks. With this update, OpenAI is proactively addressing a key vulnerability, and it will be important to watch how the implementation of Lockdown Mode impacts the overall security of the ChatGPT platform.
Sources
Back to AIPULSEN