A System for A.I. Safety

I have been researching and refining my skills over the years to build actionable A.I. Safety solutions for the industry. Well, there are multiple birds here, so bear with me. First and foremost, I have always wanted to build a wearable. I soon realized A.I. and physical human-computer interaction concepts like wearables are the true endpoint. While thinking about this bridge, I noticed wearables are the ultimate form of safety implementation, protecting the person using intelligence — "wear intelligence" being the first desired tagline of Rao.

Starting a hardware company is difficult though. It requires either having made smarter decisions early on to accumulate the capital to invest personally, or convincing a room of investors to trust you to deliver a project without the credentials to do so. If neither, it's best to make hardware the final approach or vision for a company and start with software, or with something that has less overhead. Besides, A.I. Safety starts with the software and systems first, so these needed to be built anywho, and building them to attain the capital necessary for the wearable endgame I clearly foresee makes the most sense. That's how Rao Studios started, with its initial offerings of Ambient, Veil, and Craft.

This series covers how I thought of each component A.I. Safety requires, and how to combine them into a single solution that covers most, if not all, vectors of attack, with clear signals of where liability should be placed. Working as a contractor at Apple, Inc. in 2025 gave me a lot of perspective at an industry level: where clear safety foresight is still lacking across the field, and what resources are in place to address it, if any.

I found the experience highly valuable, as a ground-truth comparison at a world-class firm is paramount to fully understanding whether the approach I am taking is proper and industrially sound. So... in 2026, I built the solutions during my own time.

This will be a 5 part series.

1. Media Safety via Veil, Veil8, VeilGarde — coming soon
2. Text & Code Safety via AII, Braid — coming soon
3. Harnessing a Reduction in Hallucinations — coming soon
4. Verification within Human in the Loop Systems — coming soon
5. 1 System, all domains, safely secured by Rao — coming soon

The first of this series will cover media, starting with images and the clear need for deepfake protection in the present day. I do believe the threat will evolve into other vectors of attack in the future, but once the groundwork is prepared and developed, it should be smooth sailing from there.