Michael Rowe

Trying to get better at getting better

Avoiding AI platform dependency by controlling your context

Google’s recent announcement of 15-month free Gemini Pro access for students is an incredibly high-value offer. Advanced AI tools (including Deep Research and NotebookLM), integration with Google Docs, and 2TB of storage is a fantastic deal, especially for students and researchers already using AI. I’ve signed up and will be making extensive use of the platform over the next year or so. However,

But here’s the problem. Fifteen months of engagement with these tools, while valuable (and to further reinforce the point, I absolutely think that this is a fantastic deal), requires a deep trade-off where you integrate your context into Google’s ecosystem. Your context isn’t just your data or documents—it’s the web of connections between your ideas, your unique patterns of thinking, and how you make sense of information. It’s how you link concept A to concept B in ways that only make sense because of your experience with concept C. This context is arguably more valuable than the content itself.

When you use AI platforms extensively, your research patterns, thinking processes, and intellectual connections become part of their ecosystem. The platform learns how you think, what connections you make, and how you approach problems. While you technically still own your data, the platform increasingly shapes how you interact with it. I suspect that, after 15 months of interacting with any platform, you’re going to be thoroughly locked in. I’m not saying that this is a bad thing, but think it’s worth pointing out.

This is different to regular tool use. One hammer is much like another, and you can switch between them easily. But when you give your context to a platform, you become dependent on it. The platform’s understanding of your context becomes a lens through which you view your own knowledge. Its suggestions, based on your past interactions, guide your thinking in ways that align with its capabilities and limitations.

Consider how you develop understanding over time. You start with a basic concept, then gradually build connections to other ideas, refine your thinking through experience, and develop nuanced views that differ from standard interpretations. This personal context—your unique way of connecting and understanding ideas—is valuable. Whilst platforms promise to enhance this process, they also shape it according to their own patterns and priorities.

Finding balance: Using AI tools while maintaining independence

The solution isn’t avoiding these powerful tools entirely, which would be counterproductive. If I haven’t made it clear, I think you definitely should be signing up for this deal if you’re eligible. However, try to think of AI tools as consultants rather than archivists, and create a workflow where all the context that emerges from your interactions with AI don’t stay in that system. Here’s a practical approach to using the the intelligence on offer through the Gemini deal, while keeping control of your context. To be clear, it’s a bit of a hassle and not a little bit inconvenient. But for me, it’s worth it, if only to have the option of walking away in 15 months when the free offer ends.

Keep your primary notes in systems you control. Use plain text files that capture your understanding and connections, not just AI-generated content. Your notes should reflect your thinking process, not what an AI thinks you should think.

Use AI tools for specific tasks but don’t rely on them to store your information. Export important work regularly into your own system. Think of AI platforms as powerful but temporary collaborators, not as the home for your intellectual development.

Maintain a simple, local-first approach to organising knowledge. This could be as basic as a folder of markdown files or as sophisticated as a personal wiki—the key is complete control. The system should adapt to your thinking, not vice versa.

Be strategic about context sharing. Not every interaction needs your full research history. Give enough context to get valuable assistance, but not so much that the tool becomes a crutch for your thinking.

This approach means missing out on some sophisticated features that come from deep platform integration. But the trade-off is worth it: you maintain control of your intellectual development whilst still benefiting from powerful AI capabilities.

Your context—your unique way of connecting ideas and building understanding—is valuable. It’s worth protecting, even if that means occasionally sacrificing convenience. Take advantage of these powerful tools, but don’t trade away something more valuable in the process: your intellectual independence.


Note: I’ve written about the idea of context sovereignty here, and am working on some ideas about what this looks like in practical terms.


Share this


Discover more from Michael Rowe

Subscribe to get the latest posts to your email.