The Bottom Line: Lockdown Mode restricts ChatGPT functions to prevent data exfiltration through prompt injection attacks and is being rolled out gradually to all user types.
OpenAI is rolling out a new Lockdown Mode for ChatGPT designed to reduce the risk of data exfiltration through prompt injection attacks. The feature is aimed at individuals and organizations working with sensitive data.
OpenAI has begun rolling out a new Lockdown Mode for ChatGPT. The feature is available for eligible personal accounts and addresses a specific security risk: data exfiltration through prompt injection attacks. This attack method exploits manipulated input prompts to trick AI models into inadvertently disclosing sensitive information.
Lockdown Mode is primarily designed for users and organizations working with confidential data and requiring stricter security guarantees. The implementation works by restricting certain ChatGPT functions that could serve as a vector for data exfiltration – such as the ability to process files or interact with external APIs. This is intended to minimize potential attack vectors that attackers could exploit to extract classified or confidential information.
For CISOs, this represents an additional control mechanism when integrating ChatGPT services into sensitive environments. Lockdown Mode is available to users with Free, Go, Plus, and Pro account types and is being rolled out in phases. Organizations should assess whether and how the feature fits into their existing data protection and security policies.
Source: thehackernews.com · Published June 6, 2026
Lumi AI News — AI-assisted curation in accordance with Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.6.5.