As ChatGPT and Gemini Become More Memorable, a Big Question Arises About Human Memory
The ability of artificial intelligence (AI) to remember user details is becoming increasingly sophisticated. Memory features on platforms like ChatGPT, Gemini, and Claude allow AI to recall preferences, habits, and even previous conversations to make interactions feel more personal and efficient. However, behind this convenience lies a deeper question: is the ability to forget actually a crucial part of healthy memory? Amanda Caswell, an AI Editor at Tom’s Guide, shared her experience after activating ChatGPT’s memory feature. The AI could remember her favourite food, her children’s names, and the last book she read. This made every conversation feel more practical, as she no longer needed to repeat the same information. Yet, this experience also sparked unease. Humans have always lived with limited memories; the brain naturally selects information it deems important, while other details slowly fade. In contrast, AI can continuously store and utilise any information a user provides as long as the memory feature is active. This trend is not limited to chatbots. Wearable devices like the Bee are designed to record daily activities for later retrieval, while Alexa+ is increasingly capable of remembering the context of previous conversations. Technology is gradually ushering humanity into an era where memory is no longer constrained by brain capacity. On the other hand, neuroscientists argue that forgetting is not a weakness. Various studies show that the ability to forget helps humans generalise information, adapt to new situations, and prevent the brain from being cluttered with irrelevant details. Even nostalgia and the process of forgiveness often depend on the fading of certain memories. This development also raises questions about privacy. The more personal information AI stores, the greater the responsibility for developers to ensure user data security. Furthermore, an ethical dilemma emerges regarding the boundary between beneficial personalisation and excessively deep information gathering. For Caswell, AI memory features are not merely an innovation to enhance convenience. This technology forces humanity to re-examine the meaning of memory itself. In the future, the challenge may not be making AI remember more, but rather determining which information should be retained—and which is better left forgotten.