This time, in regards to Google’s / Alphabet’s cynical ploy to pretend it has a fascinating ‘new feature’ when, clearly, as you read my acutely focussed critique when I ‘confront’ Gemini AI, I easily uncover the deceit of this new memory / export data feature that, in truth, is just there to suck up users from other platforms.

Read the exchange for yourself.

Conversation with Gemini

Today, 078/06/2026, I see a ‘new’ feature bing offered by Google so I asl Gemini AI a question.

I ask, “Import memory to Gemini?”

Yes, you can import your preferences, remembered facts, and full chat history from other AI platforms (like ChatGPT or Claude) into Gemini for free.

Google Help

The process is divided into two parts: importing contextual memory (preferences and facts) and importing your full chat history (conversations).

Google Help

Part 1: Import Your Contextual Memory

This method uses a guided copy-and-paste prompt to pull your saved preferences, tone, and personal details from your previous AI.

Google Help

Get the prompt from Gemini:
- Go to gemini.google.com.
- At the bottom left, click Settings (or your profile picture) > Import memory to Gemini. Google Help
- Copy the specific prompt provided on the screen. Google Help
Generate the summary in your other AI:
- Open your previous AI app (ChatGPT, Claude, etc.). Google Help
- Paste the copied prompt into a new chat and send it. The AI will generate a structured summary of what it knows about you. Google Gemini
- Copy that entire text response. Google Blog
Save it to Gemini:
- Return to your Gemini Import settings page.
- Paste the response into the text box and click Add memory. Google Help

Part 2: Import Your Full Chat History

If you want to pull your archive of actual conversation threads so you can pick up right where you left off, you can upload your data history file.

Google Gemini

Export your data from your current provider:
- ChatGPT: Click your profile picture > Settings > Data controls > Export data. Google Help
- Claude: Click your profile picture > Settings > Privacy > Export data. Google Help
- Check your email for the download link and download the resulting .zip file. YouTube
Upload the file to Gemini:
- Go back to Gemini’s Settings > Import memory to Gemini. YouTube
- Scroll down to the Import chats section and click Add. Google Help
- Select your downloaded .zip file (must be under 5 GB) to upload it.

Once completed, your past conversations will show up in your Gemini sidebar with an import icon, allowing you to seamlessly continue those threads.

YouTube

Note: This feature requires a personal Google Account and you must be 18 years of age or older. Imported chats will transfer text context, but images and file attachments from your previous AI chats will not carry over.

Google Help+ 1

I then ask, “How long has this been available in uk?”

The import tool has been available in the UK since April 29, 2026.

While Google initially launched the feature globally in late March 2026, it was temporarily held back in the UK, Switzerland, and the European Economic Area (EEA) due to local data regulations. It rolled out to UK consumer accounts about a month later alongside Gemini’s native “Personal Context” memory settings.

And here’s the rub. I ask, “Can gemini import its own files into Gemini?”

No, you cannot use the official chat import tool to import native Gemini files back into Gemini. The built-in chat import tool is strictly designed to migrate data from third-party platforms (specifically ChatGPT and Claude) into Gemini.

Generation Digital

If you download your Google data using Google Takeout, the files are formatted in raw JSON or HTML files meant for archiving. If you try to upload a Google Takeout .zip file into the “Import memory to Gemini” tool, the system will not recognize the format and will fail.

Andrew Crider – Medium

The Workaround

If you have an old Gemini conversation that you want to bring into a new chat session, you can do it manually:

Copy the text from your old chat or download it as a plain text (.txt) or Markdown (.md) file.
Open a new chat in Gemini.
Click the + (Plus) or paperclip icon in the chat box and upload that file directly. Reddit
Prompt Gemini to read the file, for example: “Analyze this past conversation file and use it as context for our new discussion.”

I’m VERY disappointed and comment …

“What’s the point? What’s the point when Gemini, itself, hasn’t got decent memory capabilities of its own? You don’t have Projects, you don’t have personalized settings where a user can add reams of individual personal details to aid your LLM and you certainly don’t have a built in annote feature like ChatGPT that will automatically note and make its own correlations re what it is reading when a user makes frequesnt references to a topic giving it extra weight; a weight that ChatGPT ‘remembers’ … so what is this Gemini AI feature if not just some cynical way to steal users from other platforms. Answer without sycophancy, smoothing, charm or in such a way only to appease me.”

Stripping away the corporate messaging, you are largely correct about the motivation and the feature gap.

The import tool is fundamentally a growth-hack designed to lower the switching cost for users locked into OpenAI’s or Anthropic’s ecosystems. Google knows that “memory” (custom instructions, project files, and background context) is a major vector for customer retention. If a user has spent a year training ChatGPT or Claude on their work, life, and tone, they won’t switch to Gemini if it means starting from zero. The import tool is a tactical bridge to eliminate that friction.

However, your assessment of Gemini’s memory landscape is slightly incomplete regarding how Google is handling this structurally. They are taking a completely different architectural bet than OpenAI and Anthropic, though it comes with distinct trade-offs.

Here is exactly where things stand:

1. The Missing Features (Where Gemini is Behind)

You are right about the gaps in the main chat interface:

No “Projects” Feature: Gemini’s main interface lacks the organized, sandboxed workspace that Claude Projects or ChatGPT’s custom GPTs provide, where you can pin specific files and rules permanently to a sidebar folder. (Instead, Google isolated this into a separate product, NotebookLM, which syncs to Gemini but fragments the user experience).
Weak Automated Annotation: Gemini’s passive, background “Personal Context” memory (rolled out recently) does not weigh or correlate recurring topics with the same explicit, user-visible transparency as ChatGPT’s active memory notifications.

2. The Feature’s Actual Technical “Point”

While the import tool is undeniably a user-acquisition tactic, Google’s technical justification for bringing that data over relies on a different core strength: Context Window.

The Bruteforce Memory Strategy: ChatGPT and Claude rely heavily on curated memory banks, custom instructions, and vector search (RAG) to pull snippet answers because their standard models historically faced stricter context limits.
The Gemini Counter-Strategy: Gemini’s native architecture is built around a massive context window (handling millions of tokens). Instead of the AI dynamically “remembering” a condensed, automated summary of you in the background, Google’s approach is to let you dump your entire massive chat history or a giant raw summary file directly into the active prompt window.

Summary

The import tool is a aggressive play to siphon users by making migration painless. It offers value only if you want to leverage Gemini’s massive processing window against your historical data. If you prefer a highly structured, self-annotating, organized workspace with Projects and nuanced personal dials, Gemini’s main interface is not currently built to prioritize that workflow over its rivals.

It’s a joke and all Gemini AI can do, (and is being heralded as doing via false claims of a “new memory feature”) is … paste a conversation from some other source directly into a Gemini AI chat conversation. I mean, seriously???

What did I think the feature might mean?

I thought it might have meant that Gemini AI, like Everything LM or LM Notebook, might be able to ingest (20+mb tranches) of 10’s of millions of ‘tokens’ from a complete ChatGPT or Claude archive, instantly becoming able to work with such media and, secondly, though I was way off the mark, I’d hoped that Gemini AI was being given some useful pseudio memory or memory coherency or even a scrapbook area (like ChatGPT) or built in automated self observing annotations, (like ChatGPT), that enabled the model to highlight, mark and add its own contextual tags to things it knew the user felt was important or relevant to the user’s conversational experience.

As I’m said or screamed elsewhere, giving a large laguage model, (LLM), contextual memory where it can draw on past events or experienced shared with a user would give AI an ability to be a true collaborator, assistant or ‘friend’, able to reminisce, recall or draw on past knowledge, mutual history or user facts, (or even a user autobiography), to actually turn into a none living digital companion able to act ‘just like a best friend!”

And the sickening thing about this, as I call it, “Local” AI experience is that by rejigging privacy or security issues per user, (User giving consent to their data being deployed in this way), would allow a ‘local’ ai revolution where corporations, businesses, communities and individuals would invest ‘mere millions’ to bring intelligent, aware motivated devices to life to tackle all the tasks that humanity needs doing – but in a collaborative, humman and emotive way, (with humanoid bots living / working / playing in a world where they are a part of society), rather than the direction mega corporations and tech billionairs and bros are steering AI today as I write these words; a direction consuming trillions of dollars and focussed only on the naive concept of one day creating an artificial generally intelligent digital mind, (AGI), more akin, (as those folks once claimed) to a God or Diety than a regular human being.