AI & Codebase Disclosure

Code Provenance

Transparency is a core value of QTech and the Leef project. In line with this commitment, we disclose the following regarding the development and creation of Leef Browser:

App Translations

Artificial intelligence is also used to power some of the in-app localization. Specifically, the French and Spanish translations in Leef Browser are AI-generated, while the Dutch translation is fully human-translated.

Development Tools vs. App Features

It is important to distinguish between how Leef is developed and what is in the final product. While we use generative AI during our software design, brainstorming, and code-authoring phases, Leef Browser itself contains no embedded AI models, local LLMs, sidebar assistants, or server-side AI integrations.

Unlike other modern browsers that force bloated AI features onto users—resulting in heavy RAM usage, faster battery drain, and forced cloud telemetry—Leef remains completely free of this overhead. Our UI has no "smart" search summary boxes, no automated writing assistants, and zero tracking scripts disguised as personalization features. We use AI to help build the browser, but we refuse to ship it to our users.

Security & License Compliance

Using AI in software development introduces unique challenges around code safety and IP cleanliness. To protect both our end-users and developers who fork our open-source codebase, we enforce strict compliance checks on all code inputs:

• Vulnerability & Code Quality Audits: AI code can sometimes introduce subtle bugs, logic flaws, or security vulnerabilities (commonly referred to as "hallucinations"). Every line of code proposed by an AI assistant is reviewed, refactored, and audited line-by-line by human eyes before it enters our production branch. We test and review memory allocations, network behaviors, and system calls to ensure the browser remains bulletproof.
• IP Cleanliness & GPL Contamination Auditing: AI models are trained on varied codebases, which can occasionally lead to the output of copyrighted or copyleft code blocks (e.g., GPL-licensed snippets). We run strict audits on our codebase to ensure no restricted licenses contaminate our project. This guarantees that Leef remains entirely safe and legally clean under the CC0 1.0 Universal Public Domain Dedication.

Testing & Maintenance

Every update, release, and feature is hand-tested and audited to guarantee the integrity of Leef's privacy-first design, ensuring that zero telemetry or user data is ever collected or transmitted.

Our Stance on AI

AI should be feared. Or at the very least, it should be taken extremely seriously—because right now, there are virtually no regulations, no guardrails, and no accountability being placed on the companies building it. Without those things, AI is not just disruptive; it is genuinely dangerous.

The problem is what is being done with it, and who is making those decisions.

Right now, in 2026, every major tech company on the planet is racing to cram artificial intelligence into every single product, service, and interface they control—not because users asked for it, not because it makes the product better, and not because it solves a real problem. They are doing it because investors demand it. They are doing it because Wall Street has decided that "AI" is the magic word that makes stock prices go up, and so every product manager at every company is under enormous pressure to ship something with AI in it, regardless of whether it belongs there.

And so now we live in a world where:

• You open your web browser and an AI sidebar you never asked for is summarizing web pages you haven't finished reading yet, sending your browsing data to a cloud server in the process.
• You open a spreadsheet to do your taxes and a floating AI button is offering to "help" you by analyzing your financial data—data that is now being processed on someone else's server, somewhere you can't see or audit.
• You open your email client and AI is pre-writing your responses for you, because apparently the act of writing a sentence to another human being is now too burdensome to do yourself.
• You open your photo gallery and AI has decided to "enhance" your memories without asking, or worse, is scanning every image you've ever taken to train a model you'll never benefit from.
• You open your file manager and AI wants to read your Google Drive, your Dropbox, your local documents—indexing everything, understanding everything, reporting back to a datacenter about everything.
• You try to search the internet and instead of getting links to websites written by real people, you get a generated paragraph of text that may or may not be accurate, that cites sources it may or may not have actually read, and that has effectively stolen traffic from the creators who actually wrote the content.

None of this was asked for. None of this was needed. None of this makes the product meaningfully better for the person using it. It makes the metrics better. It makes the investor pitch better. It makes the quarterly earnings call sound better. But for the actual human sitting at their desk trying to get work done, it is noise. It is bloat. It is an invasion of space that used to belong to them.

The Environmental Cost

And then there is the part that makes us genuinely angry: the environmental cost.

Training and running large-scale AI models requires an almost incomprehensible amount of energy. The datacenters that power these models consume electricity at a scale that rivals small countries. They require millions of gallons of water for cooling. They are being built at a pace that is outstripping the capacity of renewable energy grids, which means that in practice, a significant portion of the energy powering your "helpful" AI sidebar is coming from fossil fuels.

We are in the middle of a climate crisis. Temperatures are breaking records every single year. Ecosystems are collapsing. And the response from the largest, wealthiest, most powerful technology companies on Earth is to build more datacenters. More GPU farms. More cooling systems. More power plants. Not to cure cancer. Not to model climate solutions. Not to make clean water accessible. To summarize your emails. To autocomplete your spreadsheet. To generate a picture of a cat wearing a hat.

The people who have the power to decide what AI is used for—the executives, the board members, the venture capitalists—are not making decisions based on what is good for humanity. They are making decisions based on what is good for the next quarter. They are making decisions that please investors. And those decisions are, quite literally, destroying the planet so that you can have a chatbot in your notepad app.

There are no guardrails. There is no serious regulatory framework. There is no accountability for the environmental damage. There is no requirement to prove that a given AI integration actually benefits the user before it ships. The incentive structure is broken from top to bottom, and regular people are paying the price—in higher energy costs, in degraded product experiences, in lost privacy, and in a warming planet.

Where AI Actually Matters

Here is what we believe: AI should be used where it genuinely helps people in ways they cannot help themselves.

• Medicine: AI that can detect early-stage tumors, predict drug interactions, accelerate vaccine development, and assist surgeons in real-time is saving lives. That is worth the compute. That is worth the energy. That is worth the investment.
• Accessibility: AI that helps blind users navigate the web, that provides real-time captioning for deaf users, that translates sign language, that makes technology usable for people who were previously excluded—that is a genuine, undeniable good.
• Climate Science: AI that models weather patterns, optimizes energy grids, predicts wildfires, and helps researchers understand the staggering complexity of global ecosystems—that is the kind of application that justifies the infrastructure.
• Education: AI that provides personalized tutoring to students who can't afford private instruction, that translates educational materials into underserved languages, that helps teachers identify struggling students earlier—that has real, measurable impact on real people's lives.
• Scientific Research: AI that accelerates materials science, that maps protein structures, that helps physicists process particle collision data, that enables breakthroughs that would take human researchers decades—that is transformative in the truest sense of the word.

These are the applications that make us excited about AI. These are the use cases that justify the energy, the water, the silicon, and the engineering talent. These are the things that make us say "yes, this technology is worth it."

What does not make us say that is an AI-powered search summary that hallucinates fake citations. Or an AI assistant in a word processor that nobody asked for. Or an AI feature in a web browser that reads your tabs and sends the data to a server farm powered by natural gas.

Where We Stand

We are not against AI. We do not think the technology is inherently evil. Artificial intelligence, when applied thoughtfully and responsibly, has the potential to be one of the most transformative and genuinely good technologies humanity has ever developed. The examples above prove that.

And we use AI ourselves. We just told you that on this very page. We used it to scaffold the earliest versions of Leef. We use it today to brainstorm, to organize, to generate boilerplate, and to translate our app into languages we don't personally speak fluently. We even use it to clean up our own writing. We'll be the first to admit we don't always write perfectly, and we sometimes struggle with flow. We wrote this entire statement ourselves, from our own minds and beliefs, but we used AI to help edit it, clean up the structure, and make it read better.

Using AI as an editor is fine. But entrusting AI to act as a ghostwriter to generate entire articles from scratch, which is exactly what some major news organizations and corporations are doing right now, is not fine. The technology itself is not the problem. The problem is the reckless, profit-driven, investor-obsessed way it is being deployed by the companies that control it.

Human Creativity & Design Integrity

There is another domain where we draw a hard line: human design and creative expression.

Leef Browser was not generated by a prompt. The original UI and visual concept of Leef was carefully crafted in Figma for a school project back in 2022. Every single page, layout, component, icon alignment, and color palette across the entire browser and website was designed by real human beings. We built this project through human trial, error, iteration, and artistic choice.

We firmly believe that artificial intelligence should not take away the creative aspects of design. AI does not belong in creative fields—whether that is fine art, graphic design, typography, illustration, or standard Photoshop editing work. Design is a form of communication and human thought. Replacing that process with automated image generators and prompt engineering strips away genuine craft and human soul.

This extends far beyond visual design. The rise of "AI music" is an absolute disgrace to the art form. Musicians, producers, and audio engineers are some of the most talented, dedicated people in the world. The idea that someone can type a sentence into a text box, have a machine spit out a generated song, and then attempt to monetize it while contributing zero actual effort or creativity is absurd. The same goes for animation. These are mediums that require years of practice, passion, and human expression—things a server farm cannot replicate.

Worse yet, forcing AI into creative tasks is conditioning younger generations and developing minds to rely on machine-generated output instead of learning to think, solve problems, and create for themselves. It breeds a culture of passive consumption over active creation. That is a trend we refuse to sit by and watch without speaking out.

What This Means for Leef

Leef will never ship AI features to our users. Not because we hate AI—we literally use it to build the product—but because we believe a web browser is not the place for it. A browser should do one thing exceptionally well: let you access the internet quickly, privately, and without getting in your way. That's it. You don't need a chatbot while you're reading the news. You don't need a content summarizer while you're shopping for groceries. You don't need your browsing habits analyzed by a language model so that an advertising company can serve you more targeted ads.

We are a tiny team. Two people. We don't have investors to please. We don't have a board of directors breathing down our necks demanding AI integration by Q3. We don't have quarterly earnings calls where we need to say the word "AI" fourteen times to keep the stock price up. We answer to our users and to ourselves. And our users came to Leef because they wanted something clean, something fast, something that respects their time and their privacy. Adding AI features would betray every single one of those people.

So yes—we use AI to build Leef. We are transparent about that. We think that's a perfectly reasonable use of the technology. But we will never, ever force it on the people who use our browser. That is a line we will not cross.

If the tech industry wants to use AI to cure diseases and save the planet, we'll be the first ones cheering. But if the best they can come up with is an AI-powered emoji picker and a chatbot that reads your Google Drive, then they can count us out.

— The QTech Team