Painful intelligence:
What AI can tell us about
human suffering
Aapo Hyvärinen
University of Helsinki

Version 1.01
23rd June 2022

This is the HTML version of the e-book, see book home page for more

This book uses the modern theory of artificial intelligence (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.


1 Introduction
 Investigating intelligence by constructing it
 Is the brain a big computer?
 Machine learning as analogue to evolution
 Can an AI actually suffer?
 Intelligence is painful—overview of this book
I  Suffering as error signalling
2 Defining suffering
 Medical definitions of pain
 Medical and psychological definitions suffering
 Ancient philosophical approaches to suffering
 Two main kinds of suffering
 Using the pain system for broadcasting errors
3 Frustration due to failed plan
 Agents, states, and goals
 Planning action sequences, and its great difficulty
 Frustration as not reaching planned goal
 Defining desire as a goal-suggesting mechanism
 Intention as commitment to a goal
 Heuristics can help in planning
4 Machine learning as minimization of errors
 Neurons and neural networks
 Finding the right function by learning
 Learning as minimization of errors
 Gradient optimization vs. evolution
 Learning associations by Hebbian rule
 Logic and symbols as an alternative approach
 Emergence of unexpected behaviour
5 Frustration due to reward prediction error
 Maximizing rewards instead of reaching goals
 Learning to plan using state-values and action-values
 Frustration as reward loss and prediction error
 Expectations or predictions are crucial for frustration
 Unexpected implications of state-value computation
 Evolutionary rewards as obsessions
 Reward maximization is insatiable
6 Suffering due to self-needs
 Self as long-term performance evaluation
 Self as self-preservation and survival
 Self as desires based on internal rewards
 Uncertainty, unpredictability, and uncontrollability as internal frustration
 Fear, threat, and frustration
7 Fast and slow intelligence and their problems
 Fast and automated vs. slow and deliberative
 Neural network learning is slow, data-hungry, and inflexible
 Using planning and habits together
 Advantages of categories and symbols
 Categorization is fuzzy, uncertain, and arbitrary
 The many faces of frustration: Summarizing the mechanisms of suffering
II  Origins of suffering: uncontrollability and uncertainty
8 Emotions and desires as interrupts
 Computation is one aspect of emotions
 Emotions interrupt ongoing processing
 Desire as an emotion and interrupt
 Emtions include hard-wired action sequences
 How interrupts increase suffering
 Emotions are boundedly rational
9 Thoughts wandering by default
 Wandering thoughts and the default-mode network
 Wandering thoughts as replay and planning
 Experience replay focuses on reinforcing events
 Replay exists in rats, humans, and machines
 Wandering thoughts multiply suffering
10 Perception as construction of the world
 Vision only seems to be effortless and certain
 Perception as unconscious inference
 Prior information can be learned
 Illusions as inference that goes wrong
 Attention as input selection
 Subjectivity and context-dependence of perception
 Reward loss as mere percept
 Ancient philosophers on perception
11 Distributed processing and no-self philosophy
 Are you really in control?
 Necessity of parallel and distributed processing
 Central executive and society of mind
 Control as mere percept of functionality
 Philosophy of no-self and no-doer
12 Consciousness as the ultimate illusion
 Information processing vs. subjective experience
 The computational function of human consciousness
 The origin of conscious experience
 Why is simulated suffering conscious?
 Self vs. consciousness
 Nothing is real?
III  Liberation from suffering
13 Overview of the causes and mechanisms
 Why there is (so much) suffering
 Cognitive dynamics leading to suffering
 An equation to compute frustration
14 Reprogramming the brain to reduce suffering
 Reducing expectation of rewards
 Reducing certainty attributed to perception and concepts
 Reducing self-needs
 Reducing desire and aversion
 How far should reducing desires and expectations go?
15 Retraining neural networks by meditation
 Contemplation as active replay
 Mindfulness meditation as training from a new data set
 Speeding up the training
 Reducing interrupting desires
 Emptying the mind and reducing simulation
 Attitude of acceptance
 Metacognition and observing the nature of mind
 Letting go and relaxation as unifying principles
16 Epilogue

Copyright ©2022 Aapo Hyvärinen. All rights reserved.
Distribution allowed as per Creative Commons Attribution-Noncommercial-NoDerivatives (CC BY-NC-ND) License.