Introducing Confidential Mode: Hardware-Level Privacy for Your AI Assessment

We've always been committed to your privacy. From day one, PAICE.work was built on Privacy by Design principles. Today, we're taking that commitment even further with Confidential Mode — a new option that provides hardware-level protection for your assessment conversations.

What Is Confidential Mode?

Confidential Mode routes your PAICE assessment through Trusted Execution Environments (TEEs) — specialized hardware that creates secure, isolated spaces for processing sensitive data. Think of it as a vault within the computer where your conversation happens. Even the cloud provider hosting the service cannot see inside.

When you enable Confidential Mode, your assessment runs on NEAR AI's private cloud infrastructure, where:

Your conversation stays private — not even PAICE or the cloud provider can access it during processing
Hardware creates the protection — this isn't just a software promise, it's enforced by the computer chips themselves
Results remain yours — only your final PAICE score leaves the secure environment, no conversation data is stored or shared

Why Does This Matter?

Standard AI services process your data on regular servers. While companies like us implement strong security practices, the data is technically accessible to system administrators, the cloud provider, and anyone who might compromise those systems. Even though we have a proprietary middleware layer that strips potential Personally Identifiable Information (PII) before any LLM sees it, and even though the conversations are never stored, we're still asking you to trust us to follow our own policies with your data. The equation looks something like this:

Your PAICE Chat submission → PAICE Middleware → LLM → PAICE Middleware → Your PAICE Chat response

And when the time comes, your full conversation is transformed into your PAICE Score™ (numbers only) and the conversation is deleted.

TEE-protected inference changes that equation. The hardware itself prevents access to the data during processing. It's not that we won't look at your conversation — it's that we very literally can't. This protection is now cryptographic and verifiable, even while you remain anonymous.

This is certainly overkill for most use cases, which is why it is not the default. But for professionals working with sensitive information, regulated industries, or anyone who simply wants the strongest possible privacy guarantees, this new Confidential Mode offers peace of mind that goes beyond trust.

How to Use It

Enabling Confidential Mode is simple:

Add ?s=confidential to your assessment URL (e.g., paice.work/individual?s=confidential)
Start your PAICE assessment
Look for the TEE Protected indicator to appear at the top of the screen
Complete your assessment knowing your conversation is hardware-protected from the start

The assessment experience remains largely the same — same conversation flow, same behavioral analysis, same five dimensions, same PAICE score™. The only difference is where and how the AI processing happens, and that's now verifiably private. Also, it may be a little slower, since the data has to travel to a different server that is configued to run in a TEE. (NOTE: there is no "undo" to return to standard mode once TEE has been enabled)

Our Partnership with NEAR

This feature is powered by NEAR AI Cloud, a decentralized AI infrastructure that prioritizes user privacy. NEAR's private inference architecture runs AI models inside TEEs, ensuring that sensitive data remains confidential throughout the entire process.

We're proud to debut this functionality at NEARCON in San Franscisco next week. As part of their Innovation Sandbox program, we've opensourced this integration on Github for inspection.

We are the first AI collaboration assessment tool to offer this level of privacy protection. As AI becomes more central to professional development and hiring, the question of "who can see my AI conversations?" becomes increasingly important. With Confidential Mode, the answer is clear: only you.

Standard Mode Isn't Going Away

Our standard assessment mode continues to work securely by policy exactly as before. It uses leading AI models from Anthropic, Google, and OpenAI, optimized for the fastest and highest-quality assessment experience. For most users, standard mode provides excellent privacy (remember, we still don't store your conversations and don't sell your data).

Confidential Mode is an additional option for those who want or need the strongest possible guarantees. Choose whichever mode fits your situation.

Looking Ahead

This release marks a significant step in our mission to make AI skills assessment both insightful and trustworthy. We believe that as AI tools become ubiquitous in the workplace, users should have options for how their data is handled — options backed by technology, not just policies.

Have questions about Confidential Mode? Contact us or explore our privacy documentation for more details.

Ready to try the most private AI assessment available? Take the PAICE assessment — and enable Confidential Mode for hardware-level protection.

Get Involved:

Take the assessment (free, always)
Explore the whitepaper (comprehensive framework)
Review our privacy approach (what we collect and what we do not)
Contact us about your specific requirements (enterprise, regulated work, or privacy questions)