A multi-institutional team led by Elias Bareinboim has secured a $5 million grant from the National Science Foundation (NSF) to revolutionize artificial intelligence (AI). With our world moving closer to an AI-based economy, the delegation of decision-making to automated systems is becoming increasingly prevalent. However, these systems often exhibit societal biases and discriminatory behavior based on sensitive attributes such as gender, race, and religion. Recognizing the significance of algorithmic fairness, the team aims to build more efficient, explainable, and transparent decision-support systems to address these challenges.
Columbia Engineering is at the forefront of this transformative project, which seeks to integrate traditional AI decision-making with causal modeling techniques. By leveraging the principles of structural causal models, the team intends to create AI systems that can effectively communicate with humans and adapt to unforeseen circumstances. The project aims to unlock the potential of AI decision-making by incorporating causality and scientific methodologies into the scalability of modern AI methods.
The current generation of AI systems predominantly relies on data and statistical algorithms, but they often fall short in accurately predicting outcomes when faced with environmental changes or external interventions. Understanding the complex web of causal mechanisms underlying the environment is crucial for improving the decision-making capabilities of AI systems. With researchers across multiple institutions, including the University of Massachusetts, Amherst; University of Southern California; University of California, Irvine; University of California, Los Angeles, and Iowa State University, the team led by Elias Bareinboim will pioneer the next generation of AI systems.
The project’s real-world applications will be focused on public health and robotics. Collaborating with the Mailman School of Public Health, the team aims to enhance personalized and precise interventions for individuals with mental illness. Additionally, they will tackle the challenges of robot navigation in complex environments that involve multiple agents, such as self-driving vehicles, drones, and service robots. By developing a mobile object manipulator, the team aims to study the complexities of incorporating robot autonomy amidst humans, ensuring safety, enablement, and helpfulness.
The growing concern surrounding automation and the potential influence of AI on society forms the core motivation behind this project. By unraveling the underlying principles of AI decision-making, the team led by Bareinboim seeks to provide a scientific and causal understanding to avoid leaving crucial decisions in the hands of opaque black-box systems. The goal is to create a new generation of AI tools that align with causal principles, resulting in autonomous agents and decision-support systems that better communicate, prioritize safety, and engender trust.
Bareinboim emphasizes the significance of this research: We humans understand the world through causal lenses… If we can create AI systems aligned with causal principles, then we will be making a major advance in building a new generation of powerful AI tools for developing autonomous agents and decision-support systems that will communicate with humans, be safer, and more trustworthy.
With an aim to advance the science of causal artificial intelligence, this NSF-funded initiative signifies a remarkable stride towards ensuring AI systems are transparent, explainable, and built on a foundation of scientific understanding. By integrating causal principles with AI, the team led by Elias Bareinboim is poised to reshape the future of AI decision-making and navigate the ever-expanding frontiers of an AI-based economy.
In a world where AI is increasingly pervasive, the need for decision-making systems that are fair and unbiased has never been greater. Elias Bareinboim’s groundbreaking project holds the promise of transforming AI, ensuring that it aligns with causal principles, reflects real-world complexities, and elevates the potential of AI as a force for positive change in our society.
