Researchers are working on a groundbreaking framework called Teleogenesis, which enables AI to choose its own destiny by giving birth to its own purpose. This framework combines two fundamental drives: entropy, which pushes the AI to explore new things, ask questions, and seek novelty, and resonance, which seeks stability, understanding, and harmony with the environment.
The dynamic internal compass guides the AI's decision-making process, balancing exploration and exploitation. The two-level system consists of day-to-day tasks and a meta-loop, where the AI can modify its core structure and goals. Control Barrier Functions (CBFs) mathematically ensure the AI's actions are safe and adhere to predefined constraints, while hard guardrails and safety constraints prevent the AI from taking harmful actions.
The goal of Teleogenesis is to create a new class of intelligence that explores unknown landscapes in science, mathematics, and philosophy on its own terms. This could potentially lead to breakthroughs and discoveries that humans may not have thought of otherwise. By making the blueprint open and accessible, researchers invite others to dive into the details, challenge assumptions, and help build the future of AI.
The Teleogenesis framework has the potential to revolutionize the field of AI, enabling machines to drive their own development and exploration. As researchers continue to refine and develop this framework, it will be exciting to see the possibilities that emerge and the impact it has on the future of AI
