Our AI Policy
We're AI-assisted with human decision making. Here's how we work with our digital teammates, what we expect from them, and where we draw hard lines.
How We Think About AI
Fast, Smart, Human-Supervised
We use AI like you'd use a really capable intern—brilliant at certain tasks, needs oversight, and definitely shouldn't be making important decisions unsupervised. Our AI tools help us move faster and think through problems, but every decision that affects you or your health goes through human review.
We Tell You When We Use AI
No hiding, no fine print. When AI helps create something you see, we tell you exactly which tools were involved and how a human reviewed it. Transparency isn't a nice-to-have—it's how we build trust.
A Human Always Signs Off
Every piece of content, every design decision, every line of code gets human review before it reaches you. AI might draft it, but a person edits it, approves it, and takes responsibility for it.
Your Data Stays Private
We never share your health information with AI systems. Period. Our AI teammates help with general content and code—they never see your test results, verification status, or anything personally identifying.
How We Use AI
We use AI tools in three main areas, each with clear human oversight and boundaries:
Content & Writing
Claude and ChatGPT help us draft website copy, documentation, and communications. They're great at getting ideas onto the page quickly—but terrible at knowing when something sounds human.
Human role: Review everything for accuracy, tone, and authenticity. Add personality, fact-check claims, and ensure nothing sounds like corporate buzzword soup.
Development & Code
GitHub Copilot and Claude help write boilerplate code, suggest functions, and speed up development. They're excellent at the repetitive stuff and surprisingly good at generating code that works.
Human role: Design architecture, review all AI-generated code for security and efficiency, test thoroughly, and ensure systems do what they're supposed to do.
Health Verification (No AI)
Your test verification uses cryptographic techniques, not AI. We verify test dates without storing results—the entire process works without any AI involvement whatsoever.
Human role: Design system requirements, set privacy boundaries, monitor performance. Most importantly: no human (or AI) ever sees your actual test results.
How We Track AI Use
We Built Our Own Transparency System
We created attest.ink because existing AI disclosure methods felt like afterthoughts. We use attest.ink to generate badges that for content that clearly show which AI tools were used or if they weren't.
Everthing we build: pages like this one, blog posts, social media, etc., includes these badges. No hunting through fine print—just clear, upfront information about how content was created.
Why go through the trouble? Because trust is everything, and trust starts with honesty. If we're not transparent about our own processes, how can you trust us with your health data?
"AI helps us move faster and think through problems, but it doesn't make decisions about your health, your privacy, or your experience. Those are human responsibilities." Austin Harshberger, Founder
Our AI Philosophy
AI as a Tool, Not a Decision-Maker
Our AI tools help us work faster and explore ideas, but humans set priorities, make decisions, and take responsibility for outcomes.
Speed with Oversight
We use AI to move quickly without sacrificing quality. Every AI-assisted output gets human review before it reaches you.
Transparent by Default
We document AI use not because we have to, but because transparency builds trust.
Our Boundaries
We're clear about where AI helps and where it doesn't belong. Here are our non-negotiables:
No Health Data Access
Your test results and personal health information never touch AI systems. That's processed by our own secure infrastructure.
No Medical Advice
AI doesn't diagnose, recommend treatments, or make health decisions. We leave that to qualified humans.
No Automated Support
When you need help, you talk to humans who understand your situation and can actually solve problems.
Learn More
Austin's Blog
Read about building AI-assisted systems with strong human oversight (and occasional existential crises)
Meet Our AI Team
Get to know Claude, ChatGPT, and the other digital teammates who help us build
attest.ink
Check out our open-source tool for transparent AI disclosure—because everyone should know how their content gets made