# About Akritas Systems

Akritas Systems helps organizations use capable technology without surrendering human judgment, responsibility, dignity, or ownership.

## Our measure

Efficiency, speed, accuracy, and cost matter, but none of them is a sufficient measure of progress on its own. We also ask whether a system supports human judgment, respects knowledge earned through practice, clarifies responsibility, and preserves a real path to question, correct, or stop what the technology is doing.

## Working convictions

- **Human authority:** Models can inform, organize, and assist. They do not inherit the moral authority or accountability of the people using them.
- **Dignity before efficiency:** People are never merely inputs, costs, obstacles, or sources of data to be optimized away.
- **Discernment before deployment:** We avoid both uncritical enthusiasm and reflexive fear. We examine the real workflow, consequences, and limits first.
- **Work worth doing:** Automation should reduce needless burden and help people bring more judgment, craft, attention, and care to their work.
- **Power with boundaries:** Access, data, authority, cost, and failure behavior should be constrained deliberately.
- **Shared responsibility:** Good systems are built with the people who know the work. Technology should strengthen participation, not impose uniformity from above.

## An acknowledged influence

*Magnifica Humanitas* gave fuller language to a conviction that shaped the beginning of Akritas's AI journey: technical power does not grant the right to govern the people it affects. The document is explicitly Catholic. Akritas serves people and organizations of many beliefs, and no customer is asked to adopt its theology. We do openly carry forward its insistence that human intelligence, conscience, and freedom guide technology and set its limits.

Source: [Magnifica Humanitas — The Holy See](https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html)

## In practice

These convictions lead to approved sources, explicit permissions, visible uncertainty, human decision points, protected credentials, finite limits, staged deployment, evaluation, monitoring, and rollback.

**Build the capability. Preserve the person.**