- November 22, 2026
- Updated 8:26 am
OpenAI’s New Security Feature and AI Safety Concerns
- 37 Views
- admin
- November 2, 2026
- Tech Companies Technology
OpenAI Addresses AI Safety and Privacy Concerns
Chris Lehane, OpenAI’s Chief Global Affairs Officer, recently spoke about an investigation involving a hack of an Australian government Medicare portal. Lehane provided context for the security incident that occurred in July and updated findings from their ongoing review. Additionally, a Fox News poll indicating that 71 percent of Americans oppose local AI data centers was discussed during his interview with Lawrence Jones.
Introduction of Locked Chats Feature
OpenAI has introduced a new feature called Locked Chats allowing users to protect individual ChatGPT conversations with a four-digit PIN. Supported devices also enable the use of biometric authentication like Face ID or Touch ID. This feature offers reassurance for those seeking advice through AI while maintaining privacy.
The Locked Chats do not provide end-to-end encryption but offer convenience and control over visible conversations. Privacy limitations should be noted before locking conversations.
How to Use Locked Chats
To lock a conversation:
- Open ChatGPT and sign in to your account.
- Find the conversation to protect.
- Open the conversation’s menu and select ‘Lock chat.’
- Create a four-digit PIN to confirm.
Your conversation moves to ‘Locked chats’ under ‘Projects.’ Use device authentication like Face ID on iPhones or equivalent authentication on Android devices for convenience.
Privacy Features and Limitations
Although Locked Chats add a layer of privacy, they do not include end-to-end encryption. Locked Chats remain visible momentarily on a device until refreshed. Shared links stay active unless revoked, which necessitates separate deletion.
Organizational policies still apply to workplace accounts, impacting compliance. Locked Chats should not be used for sensitive information like passwords or Social Security details.
Preventing AI Training with Your Conversations
You can disable ChatGPT’s use of conversations for AI model training.
- Open ChatGPT settings and navigate to ‘Data controls.’
- Disable ‘Improve the model for everyone’ to stop their use for training models.
Locked Chats vs. Temporary Chats
Temporary Chats differ from Locked Chats by not appearing in regular history and by offering short-term visibility without retaining memories for AI training. The right choice depends on whether you wish to maintain long-term privacy or keep a temporary discussion.
PIN Recovery Limitations
Forgetting your ChatGPT PIN results in the permanent deletion of all protected content. Conversely, PIN changes while remembering the current PIN will retain protected conversations.
Kurt’s Insights
The introduction of Locked Chats provides users more control over their interactions with AI. As artificial intelligence becomes ingrained in personal decision-making, the ability to secure privacy becomes paramount. Despite the improvement, Locked Chats remain a convenience feature, lacking the confidentiality guarantee.
Whether OpenAI will offer more robust privacy measures is yet uncertain.