Aridio Silva

LoRA in AI Agents: Efficiency, Security Risks, and Adapter Governance

Low-Rank Adaptation (LoRA) made it possible to specialize large models by training a fraction of their parameterswhile keeping the base model frozen. This efficiency fostered an ecosystem of small, interchangeable,dynamically served adapters, but it also created a new trust objec...

When AI Agents Form Societies: Is Security a Property of the Model or the System?

Evaluating a language model on a bounded prompt tells us something about its responses under those conditions. It tells us less about a persistent group of agents that remembers previous interactions, uses tools, exchanges messages, and changes a shared environment. Two Emergence...

SGAEIA: A Reference Architecture for Secure Governed Autonomous Edge Intelligence

Artificial intelligence becomes an architectural security problem when a model can do more than generate content. Tool use, persistent memory, delegated tasks, code execution, credentials, network access, and physical actuation allow model-generated intentions to cross technical...

Ten Days That Exposed the Governance Gap in Frontier AI

For most of the modern generative-AI era, progress has been described primarily through capability: stronger reasoning, better coding, longer context windows, improved multimodal understanding, more effective tool use, and increasingly autonomous behavior. In September 2026, howe...

Ten Days That Exposed the Governance Gap in Frontier AI

For most of the modern generative-AI era, progress has been described primarily through capability: stronger reasoning, better coding, longer context windows, improved multimodal understanding, more effective tool use, and increasingly autonomous behavior. In September 2026, howe...

When AI Begins to Build AI

When AI Begins to Build AI examines the emerging intersection between AI-assisted AI research, autonomous agents, recursive self-improvement (RSI), AI Security, and frontier AI governance. The article analyzes recent developments in which increasingly capable AI systems are parti...

Evidence-as-Code: Making AI Governance Verifiable

As autonomous AI systems move from recommendation to execution, governance can no longer remain a collection of policy documents, periodic assessments, and retrospective audit reports. An agent may select tools, access resources, delegate subtasks, modify external systems, and ac...

Evidence-as-Code: Making AI Governance Verifiable

As autonomous AI systems move from recommendation to execution, governance can no longer remain a collection of policy documents, periodic assessments, and retrospective audit reports. An agent may select tools, access resources, delegate subtasks, modify external systems, and ac...

Continuous GRC for Agentic AI

This article examines Continuous Governance, Risk, and Compliance (Continuous GRC) as a foundational requirement for autonomous and agentic AI systems. It argues that periodic assessments and retrospective audits are insufficient for systems that continuously interpret goals, del...