Google adds memory to Gemini AI for personalised conversations
The new Gemini AI memory feature builds on past chats, rather than starting fresh each time, aiming to make conversations more personal.
Google has introduced an opt-in memory feature for its Gemini AI chatbot, allowing it to remember user details such as dietary preferences or travel interests for more personalised replies.
The 13 August update means conversations can build on earlier chats instead of starting from scratch, with the rollout beginning immediately for users of the Gemini app. The change positions Google in direct competition with ChatGPT by prioritising customisation to strengthen user loyalty.
The memory function is paired with privacy tools, including a Temporary Chat mode, which prevents saving conversations, and a data editing interface to view, amend, or delete stored details.
Google emphasises that users must choose to activate memory instead of having it switched on automatically.
However, concerns have emerged over a requirement, starting 2 September, for users to opt out if they do not want a sample of their chats and files used to train AI models.
Privacy advocates warn the default opt-out approach risks eroding trust, even as Google promotes safeguards like on-device processing via Gemini Nano for Android.
Regulators, particularly under the EU’s AI Act, will likely scrutinise the balance between innovation and privacy.
Industry analysts suggest that the update reflects a broader push for AI systems to be more user-centred while retaining strict data control.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
