Learning from biological and social systems for designing beneficial AI
Principles of Intelligent Behavior in Biological and Social Systems (PIBBSS) is a research initiative aiming to leverage insights on the parallels between intelligent behaviour in natural and artificial systems towards progress on important questions in AI risk, governance and safety.
We run a number of programs to facilitate this type of research, support talent and build a strong research network around this epistemic approach.
- The Summer Research Fellowship is a 3-months program, pairing fellows from disciplines studying complex and intelligent behaviour in natural and social systems, with mentors from AI alignment.
- Our Affiliate Program aims to provide tailored, long-term support to promising researchers with a strong track record pursuing “PIBBSS-style” AI alignment research
- Our semi-regular Speaker Events bring together researchers from both within and adjacent to AI alignment, and create a space to explore the connections between different lines of research as well as their relevance for AI alignment.
We are always on the lookout for other ways to support research and training in this space.
Understanding the underlying principles that govern intelligent behaviour — across systems, scales and substrates — towards safer AI futures.
Future of Artificial Intelligence
We are dedicated to prioritizing and fostering careful, value-sensitive, and epistemically diverse approaches toward the design and implementation of artificial intelligent systems.
Humanity is making fast progress in its ability to engineer sophisticated forms of complex and intelligent behaviors in artificial systems. Such systems will have an increasingly dominating effect over the future course of our civilization.
This progress could unlock immense positive potential, but it could also lead to irreversible, harmful effects. Thus, research in AI alignment aspires to contribute to preserving and fostering human potential. You can learn more about AI alignment by consulting the resources outlined here.
Intelligence in Biological and Social Systems
Insofar as AI research is interested in mechanisms that can search and instantiate complex behaviors, the study of complex natural and social systems can inform the design of artificial behaviors.
Throughout the history of science, we observed that a wide variety of complex and intelligent systems exhibit similar properties. Such cross-system phenomena include adaptation, robustness, goal-directed behavior, learning, embeddedness, modularity, phase transitions, and more.
In understanding the mechanisms that give rise to these phenomena and the role they play in defining the functionality of such systems, we seek to improve our ability to develop artificial systems that reliably depict desired properties.
PIBBSS targets a diverse set of scientific domains that have the potential to help advance key bottlenecks in designing safe and aligned artificial intelligent systems.
The fellowship is focused on domains with a rich repertoire of tools and insights about complex behavior found in natural and social systems.
These are fields such as ecology, evolutionary biology, cognitive science, neuroscience, sociology, legal theory, political economy, statistical mechanics, linguistics, media studies, and more.