AITEL Research

Research & Ideas

We explore the relationship between artificial intelligence, autonomy, memory and people.

Featured article
AI agents

An AI Agent Shouldn't Be Able to Execute Everything It Proposes

What a model proposes, what a system can execute and what has the authority to be executed are three different things. The difference between them has to be built into the architecture, not just requested in instructions.

A model can propose an action, and a system can have the technical capability to carry it out. But the authority to execute must be granted separately and enforced by a control layer that the model does not control on its own.

16 min read

Recent articles

AI safety

AI Doesn't Have to Turn Evil to Become Dangerous

Why AI risk depends not on hostile intent but on the combination of persistent memory, autonomy, access to tools and speed — and on people's right to be able to say “no.”

12 min read
Themes

What we write about

  • AI safety

    Human control, limits, permissions and the ability to say no.

  • A.D.E.

    The architecture of Autonomous Digital Entities.

  • Persistent memory

    Memory and continuity in AI systems.

  • AI agents

    AI systems that act — with tools, autonomy and consequences.

  • Human–AI relationship

    How people and AI systems relate over time.

  • AI & society

    Dependence, delegation and the impact of AI on society.

In preparation

Coming next

These articles are being prepared. Titles may change before publication.

  • AI agents

    When AI Starts Making Decisions in Our Place

    Coming soon
  • AI safety

    Can the AI That Protects You End Up Controlling You?

    Coming soon
  • AI & society

    AI Doesn't Have to Conquer Us for Us to Become Dependent on It

    Coming soon
  • Human–AI relationship

    What Does Human Control Mean in a World of AI Agents?

    Coming soon