Arr, me hearties! Avast ye! ChatGPT be gainin' a memory akin to a human, mark me words! 'Tis a proper leap toward General AI, mateys!
2024-02-13
Arrr, methinks ChatGPT be teachin' itself the art o' rememberin' yer tales, whether 'tis 'bout yerself, yer likin's, or yer fancy. Aye, a clever matey it be!
OpenAI has announced a limited test for ChatGPT, their generative AI, to remember what users tell it across all chats. This new feature allows ChatGPT to remember casual bits of information and explicit instructions, making it more like a helpful assistant who knows your preferences. The memory feature will be applied to future prompts, allowing ChatGPT to provide more personalized and efficient responses.Some users might find the idea of an AI that remembers multiple conversations off-putting, so OpenAI is offering an opt-out option. Users can choose the "Temporary Chat" mode to introduce amnesia to ChatGPT or manually delete memories through the settings. However, OpenAI has promised that memories will be excluded from ChatGPT's training data, ensuring privacy.
OpenAI is also adding memory capabilities to their new GPTs, allowing developers to create bespoke chat-based AIs. However, these developers will not have access to the memories stored within the GPT.
This update could have significant implications, as memories can enhance the perception of ChatGPT's human-like qualities. Shared experiences and memories are crucial in human relationships, and the ability for ChatGPT to remember details and preferences may create a stronger connection with users.
While the memory feature is currently being tested among some users, there is no timeline for when it will be available to all users. OpenAI is also taking steps to mitigate biases and prevent ChatGPT from proactively remembering sensitive information unless explicitly requested.
Overall, the introduction of memory in ChatGPT marks a small but significant step towards making AI more human-like and paving the way for more advanced General AI systems in the future.