Agentic X

The notion of a 𝘀𝗽𝗲𝗰𝘁𝗿𝘂𝗺 𝗼𝗳 𝗮𝗴𝗲𝗻𝗰𝘆 with 𝘃𝗮𝗿𝘆𝗶𝗻𝗴 𝗱𝗲𝗴𝗿𝗲𝗲𝘀 𝗼𝗳 𝗮𝘂𝘁𝗼𝗻𝗼𝗺𝘆 and imbedded in applications gives rise to something I like to refer to as 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗫

The notion of a 𝘀𝗽𝗲𝗰𝘁𝗿𝘂𝗺 𝗼𝗳 𝗮𝗴𝗲𝗻𝗰𝘆 with 𝘃𝗮𝗿𝘆𝗶𝗻𝗴 𝗱𝗲𝗴𝗿𝗲𝗲𝘀 𝗼𝗳 𝗮𝘂𝘁𝗼𝗻𝗼𝗺𝘆 and imbedded in applications gives rise to something I like to refer to as 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗫. Where applications and interfaces are imbued with different levels of agency.

AI Agents are broadly defined as software entities capable of perceiving and acting upon their environment.

However, in the era of 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 (𝗟𝗟𝗠𝘀), the term 𝗔𝗴𝗲𝗻𝘁 has taken on a narrower meaning, as even simple systems like thermostats would qualify as agents under the traditional definition.

Researchers have sought to formalise a shared understanding of what constitutes an 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁 𝗶𝗻 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲-𝗺𝗼𝗱𝗲𝗹-𝗯𝗮𝘀𝗲𝗱 𝘀𝘆𝘀𝘁𝗲𝗺𝘀, often framing it as a 𝘀𝗽𝗲𝗰𝘁𝗿𝘂𝗺 𝗼𝗳 “𝗮𝗴𝗲𝗻𝘁𝗶𝗰” behaviour rather than a binary state.

This perspective emphasises 𝘃𝗮𝗿𝘆𝗶𝗻𝗴 𝗱𝗲𝗴𝗿𝗲𝗲𝘀 𝗼𝗳 𝗮𝘂𝘁𝗼𝗻𝗼𝗺𝘆 & functionality within systems rather than rigid classifications.

Instead of proposing a new definition, we should rather identify factors that contribute to an AI system being perceived as more agentic.

These factors can be grouped into four main clusters as seen in the image, which together provide a nuanced framework for understanding the evolving concept of agency in AI

Chief Evangelist @ Kore.ai | I’m passionate about exploring the intersection of AI and language. From Language Models, AI Agents to Agentic Applications, Development Frameworks & Data-Centric Productivity Tools, I share insights and ideas on how these technologies are shaping the future.

https://arxiv.org/abs/2407.01502

https://www.cobusgreyling.com