ChatGPT learns to forget — and that’s exactly what we needed 🧠

Cliquez ici pour lire en français

Imagine an AI assistant that can not only remember everything you tell it—but also know what to forget. Until now, ChatGPT’s memory was limited by technical constraints, meaning it could only retain a finite amount of information. But OpenAI just rolled out a groundbreaking update: ChatGPT can now manage its own memory, automatically forgetting less relevant data to make room for what truly matters. This marks the beginning of a new era for AI interactions—smarter, smoother, and genuinely more personal.

Goodbye full memory, hello smart management 🧹

In the past, once ChatGPT’s memory filled up, it could lose track of key details or force users to repeat information. The new system fixes that. ChatGPT now sorts and prioritizes data based on how often you reference it and how important it seems to your conversations.

In other words, it remembers what actually helps it understand you—without unnecessary clutter. Even better, users can now set their own rules about what’s important or disposable. It’s a major step forward in solving that old frustration of “starting over” with your AI every few weeks.

Users take back control of AI memory 🎛️

OpenAI built transparency and flexibility into this system. You can view, reorganize, or completely delete ChatGPT’s stored memories whenever you want. No more worrying about an AI “that remembers everything without filters.” This selective forgetting is a genuine milestone—it brings us closer to AI that not only learns from you, but adapts to you. An assistant that grows alongside you, without getting bogged down by irrelevant details.

Source : © OpenAI

A premium feature — for now💎

There’s a catch: this intelligent memory system isn’t available to everyone yet. It’s currently limited to ChatGPT Plus and Pro subscribers, as part of an ongoing global rollout. Still, early access users report a major improvement in how natural and relevant conversations feel. If this feature scales globally, it could redefine how we think about long-term AI companionship—less robotic, more human.

Forgetting to remember better 🤔

By giving ChatGPT the ability to forget, OpenAI is taking a bold philosophical step. Should AI be infinite libraries—or smart partners that know what to keep and what to let go? This new memory system pushes us to rethink what it means for an AI to know us. Users now hold the power to shape their digital assistant’s personality, history, and focus. One thing’s clear: the era of “memory full” AIs might finally be coming to an end.

👉🏾 What do you think about an AI that forgets?
Is selective memory a breakthrough—or does it risk losing something important? Share your thoughts in the comments!