Google Gemini dismantles AI silos by importing user memories from ChatGPT and Claude

The landscape of generative artificial intelligence has long been characterized by walled gardens, where users who spend months refining a chatbot’s understanding of their preferences, professional context, and personal quirks find themselves tethered to a single platform by the sheer weight of their accumulated data. This barrier to entry for rival services, often referred to as platform lock-in, is now being systematically dismantled. Google has launched a significant update to its Gemini assistant that introduces a streamlined system for importing "memories" and chat histories directly from competitors like ChatGPT and Claude.[1][2][3][4][5][6] This move follows a similar initiative by Anthropic and represents a pivotal shift in the industry toward data portability and interoperable personal intelligence.

At the heart of this update is a dual-pronged approach to data migration that addresses both the high-level personal context of a user and the granular history of their past interactions.[4][7][6][1] The first and most innovative component is a prompt-based "memory" import tool.[8] For users who have utilized ChatGPT’s memory feature or customized Claude’s behavior over time, the thought of re-explaining their writing style, professional background, or specific project requirements to a new model is often enough to prevent them from switching services.[9] Google’s new tool mitigates this by providing a specialized prompt that users can copy and paste into their existing AI assistant.[1][6] This prompt instructs the rival model to analyze its stored data and generate a comprehensive summary of everything it knows about the user, including demographic details, recurring topics, and explicit instructions.

The technical execution of this prompt is designed specifically for cross-platform compatibility. It requires the source AI to output information in a neutral, third-person format, avoiding first-person pronouns like "I" or "my" and second-person pronouns like "you" or "your." By framing the output as a description of "the user," the resulting data packet is structured in a way that Gemini can instantly ingest and integrate into its own long-term memory. Once this summary is pasted into Gemini’s settings, the AI adopts the established persona and context of the user, effectively allowing the "relationship" between human and machine to continue without a reset. This effectively transforms a once-tedious manual process into a simple two-step copy-paste operation, significantly lowering the friction required to trial or switch to Google’s ecosystem.

Beyond high-level summaries, the update introduces a robust "Import Chats" feature that allows for the transfer of entire conversation histories.[10][7][4] Users can now export their data from ChatGPT or Claude as a standard ZIP file and upload it directly to Gemini.[2][6] Google supports files up to five gigabytes in size and allows users to upload multiple archives per day.[6] Once processed, these imported chats appear in the Gemini sidebar, marked with a distinct icon to differentiate them from native conversations.[6] This allows users not only to archive their previous work but to actively continue those specific threads within the Gemini interface.[11][4] While the tool currently focuses on text-based prompts and responses—excluding attachments, project files, and AI-generated images—it provides a level of continuity previously unavailable in the consumer AI market.

This strategic move by Google arrives just weeks after Anthropic pioneered a similar prompt-based migration strategy for its Claude assistant. The timing is notable, as the AI sector has seen a surge in users looking for alternatives to OpenAI following various corporate controversies and shifts in product direction. By adopting and expanding upon Anthropic’s methods, Google is signaling that it no longer views "memory" as a proprietary anchor used to keep users trapped. Instead, it is positioning Gemini as a flexible, high-capacity alternative that can meet users where they are, regardless of where they started their AI journey.[3] The renaming of Gemini’s "Past Chats" to "Memory" further reinforces this shift toward a more holistic, persistent assistant that evolves with the user across different sessions and platforms.

The implications of this update for the broader AI industry are profound. For years, the competitive advantage of an AI assistant was measured not just by the quality of its underlying model, but by the depth of its user-specific data silo. By making that data portable, Google and Anthropic are forcing a shift in competition toward model performance, feature sets, and ethical positioning. If a user can move their entire digital identity from one AI to another in under five minutes, the "stickiness" of a platform must be earned through continuous innovation rather than data hostage-taking. This could lead to an era of "AI interoperability," where a user’s personal context becomes a portable asset they own and control, rather than a commodity owned by a specific service provider.

However, this transition to open borders in the AI space is not without its hurdles.[5] Currently, Google has restricted the availability of these import tools, excluding the European Economic Area, Switzerland, and the United Kingdom from the initial rollout.[6] This is likely due to the complex regulatory environment created by the General Data Protection Regulation and the emerging EU AI Act, which place stringent requirements on how personal data is processed, transferred, and stored by large-scale AI models. The challenge for these companies moving forward will be to maintain this new level of portability while navigating the legal landscapes of different jurisdictions that prioritize data sovereignty and privacy.

Furthermore, while the data can be moved, the "personality" and utility of that data may still vary between models. A set of preferences that works perfectly for the creative, conversational tone of Claude might produce different results when applied to the more analytical and integration-heavy framework of Gemini. As users begin to migrate their data in bulk, the industry will likely see a new focus on "context translation"—the ability of an AI to not just read imported data, but to interpret and apply it in a way that feels consistent with the user's original intent. The current prompt-based solution is a clever workaround to the lack of official APIs for data transfer between rivals, but it serves as a precursor to more formal standards for AI data portability.

In conclusion, Google’s decision to simplify the migration of memories and histories marks the end of the first era of AI silos. By empowering users to take their context with them, Google is betting that Gemini’s integration with the broader Workspace ecosystem and its advancing multimodal capabilities will be enough to win over users who were previously "locked in" elsewhere. This update does more than just add a feature; it changes the social contract between AI providers and their users.[1] It moves the industry one step closer to a future where artificial intelligence is a truly personal, portable utility that serves the individual rather than the platform. As the walls between these digital assistants continue to crumble, the true winner will be the user, who finally gains the freedom to choose the best tool for the task without losing the history of their digital life.