Painful intelligence: What AI can tell us about human suffering

Abstract: This book uses the modern theory of AI to understand human suffering or mental pain. Both humans and sophisticated AI agents process information about the world in order to achieve goals and obtain rewards, which is why AI can be used as a model of the human brain and mind. This book intends to make the theory accessible to a relatively general audience, requiring only some relevant scientific background. The book starts with the assumption that suffering is mainly caused by frustration. Frustration means the failure of an agent (whether AI or human) to achieve a goal or a reward it wanted or expected. Frustration is inevitable because of the overwhelming complexity of the world, limited computational resources, and scarcity of good data. In particular, such limitations imply that an agent acting in the real world must cope with uncontrollability, unpredictability, and uncertainty, which all lead to frustration. Fundamental in such modelling is the idea of learning, or adaptation to the environment. While AI uses machine learning, humans and animals adapt by a combination of evolutionary mechanisms and ordinary learning. Even frustration is fundamentally an error signal that the system uses for learning. This book explores various aspects and limitations of learning algorithms and their implications regarding suffering. At the end of the book, the computational theory is used to derive various interventions or training methods that will reduce suffering in humans. The amount of frustration is expressed by a simple equation which indicates how it can be reduced. The ensuing interventions are very similar to those proposed by Buddhist and Stoic philosophy, and include mindfulness meditation. Therefore, this book can be interpreted as an exposition of a computational theory justifying why such philosophies and meditation reduce human suffering.

Insights into AI and Human Suffering: An Analytical Overview

Aapo Hyvärinen's work presented in "Painful Intelligence: What AI Can Tell Us About Human Suffering" provides an intriguing perspective on understanding human suffering through the lens of artificial intelligence (AI). The book explores the parallels between AI models and the human brain, offering a computational viewpoint on suffering, primarily through the concept of frustration, and delves into interventions to alleviate such suffering. This synthesis, while rooted in scientific theories, also draws philosophical parallels with Buddhist and Stoic practices, notably mindfulness meditation.

Frustration as the Core of Suffering

Hyvärinen begins by articulating a theory where human suffering is analogized to an AI's inability to achieve goals due to various computational limitations. In an AI context, frustration emerges when an agent fails to achieve expected outcomes, reflecting a core error-signalling mechanism. This analogy stems from AI agents striving to optimize goal achievement and reward acquisition, processes mirrored in human cognitive and emotional systems. The complexity and unpredictability of the environments combined with limited computational resources inherently lead to frustration, which Hyvärinen identifies as a root cause of suffering.

Cognitive Models and Computational Analogies

The majority of Hyvärinen's theoretical grounding is in the field of machine learning, particularly reinforcement learning, where an AI's objective is to maximize reward through adaptive learning. By paralleling human cognitive processes to machine learning algorithms, Hyvärinen discusses how humans learn from error signals in a manner comparable to AI. The concept of 'reward prediction error' (RPE) becomes a cornerstone, representing the discrepancy between expected and received rewards, directly linking it to experiences of frustration.

The Role of Self and Survival Instincts

Another layer involves evaluating the human experience in terms of self-needs—specifically, self-preservation and self-evaluation. In an AI framework, these translate to intrinsic motivations or internal rewards, essential for long-term survival and learning optimization. Hyvärinen suggests that these self-needs correlate with long-standing psychological constructs, such as self-esteem and depression, attributed to persistent goal failures. The analogy with AI systems is extended further to propose that a lack of reward or unmet expectations can metaphorically 'frustrate' an AI, drawing distinct parallels with human psychological phenomena.

Limitations and Extensions of Control

A significant theme is the loss of control, both in humans and AI systems, characterized by Hyvärinen's treatment of emotions and desires. Emotions—conceptualized as interrupts—effectively reduce cognitive control, analogous to errors interrupting computational processes. This framework extends the dual-process theories in psychology, where fast, unconscious processes might disrupt more deliberative, conscious processing, akin to neural network fast processing disrupting symbolic computing in AI.

Learning from Misperceptions: Uncertainty and Untapped Potential

Hyvärinen critically examines the illusion of control and the inherent uncertainty in both AI perceptions and human perception. Both systems attempt to create coherent world-models from incomplete data, yet this process is riddled with ambiguity and subjective interpretations. Such insights from AI models shed light on the limitations of human perception and cognition, drawing parallels to how AI systems predict, process, and learn.

Implications for Alleviation of Human Suffering

Perhaps one of the most profound contributions of Hyvärinen's research is using this computational framework to propose interventions akin to methods suggested by Buddhist and Stoic philosophies. The interventions target reducing frustration and suffering by employing techniques such as mindfulness, which aligns with optimizing cognitive processing and reducing error-based signals in AI terms.

Future Prospects in AI and Human Cognition

Hyvärinen's integration of computational theories of AI with human cognitive models opens numerous pathways for future research. Understanding how AI's learning mechanisms can mirror human psychological states offers an enriched perspective on alleviating suffering, promoting interdisciplinary exploration between AI, psychology, and philosophy. Moreover, this approach encourages the development of AI systems that perhaps not only mimic but also help enhance human cognitive capacities in coping with complexity and unpredictability.

Overall, "Painful Intelligence" provides a robust framework for evaluating human suffering through AI, pushing the boundaries of both fields to explore symbiotic relationships that promise both technological advancement and cognitive growth.