OpenAI is rolling out an enhanced memory feature for ChatGPT that allows the chatbot to reference past conversations across all interactions. The update aims to make ChatGPT’s responses more personalized and relevant by drawing on previous discussions without requiring users to repeat information.
The new capability will initially be available only to ChatGPT Plus and Pro subscribers, with Enterprise, Team, and Edu users gaining access later. Users in the UK, EU, Iceland, Liechtenstein, Norway, and Switzerland will need to wait for additional regulatory reviews before accessing the feature.
User controls for privacy
OpenAI has implemented two main control settings:
- “Reference Saved Memories” – ChatGPT stores specific facts or preferences
- “Reference Chat History” – The AI draws context from previous discussions
Users can disable these features at any time, manage specific saved memories, or use Temporary Chat for conversations they don’t want stored. The memory feature will be enabled by default for users who previously had ChatGPT’s memory capabilities turned on.
The expanded memory capabilities have sparked mixed reactions. While some industry experts like AI investor Allie K. Miller highlight memory as a key competitive advantage, others express concerns. Wharton professor Ethan Mollick noted he won’t enable the feature, preferring boundaries between personal and professional AI interactions. OpenAI cofounder Andrej Karpathy even worried the AI might “think worse” of him based on past questions.
Sources: TechCrunch, VentureBeat