Bots & Beasts: What Makes Machines, Animals, and People Smart? will be published by MIT Press in 2021.
Photo by Jenna Oxley
Magpies can recognize themselves in mirrors and octopuses can open jars to get at food. An IBM computer won a debate with people about the costs and benefits of artificial intelligence, and cars are beginning to drive themselves. How do computers and animals measure up to the minds of people?
Human intelligence is more than IQ thanks to many features such as decision making, learning, and creating. These features generate report cards for evaluating smart computers and animals. The ability to learn how to learn is just one of the advantages that people have over current bots and beasts. Consciousness only began when animals such as fish developed brains large enough to combine thoughts. Artificial intelligence is advancing but still falls short of the human capacities for consciousness, emotion, and creativity. These conclusions have important implications for the moral treatment of animals and intelligent computers.
This book is the first systematic comparison of intelligence across machines, humans, and other animals. It draws on philosophy for a method of attributing mental capacities to nonhumans and for an approach to ethics based on vital needs. Psychology and neuroscience furnish the mechanisms that support intelligence. The assessment of limitations in current approaches to AI provides a clear blueprint for future research to approximate human-level intelligence. Policies are recommended to ensure that such developments promote the satisfaction of human needs.
Policy Recommendations (from ch. 8)
I have systematically compared intelligence in machines, humans, and other animals by assessing their features and mechanisms. The results have implications for the study of intelligence, the treatment of animals, and the development of artificial intelligence. 50 years from now, if we have managed to survive climate disasters, pandemics, and evil leaders, we should have a much deeper understanding of all kinds of intelligence and developed new technologies for augmenting but not replacing human intelligence.
Investigation of Intelligence Should Be Interdisciplinary
The field of psychology has made many valuable contributions to understanding intelligence, but its narrow focus has had unfortunate consequences. Fixation on IQ and g has skewed the concept of intelligence towards linguistic and mathematical skills and away from valuable nonverbal aspects such as emotions and spatial navigation. Mathematical techniques that look for associations among measurable variables have discouraged the discovery of underlying mechanisms that provide deeper explanations. Concentration on IQ and correlational methods has sometimes enabled arguments intended to undermine racial equality.
The best way to broaden psychological investigation of intelligence is to combine it with ideas and methods from other fields, in line with the interdisciplinary aims of cognitive science. Artificial intelligence provides computational models of problem solving, learning, and the other features of intelligence that serve to identify mechanisms responsible for human thinking. Neuroscience also contributes to explanatory mechanisms ranging from molecular processes such as neurotransmitters to the interactions of networked brain areas. Ethology adds evidence and theories about intelligence in non-human animals. Anthropology ensures that accounts of intelligence are valid across diverse cultures. Because so much of human intelligence is language-oriented, the field of linguistics also provides insights. Finally, philosophy has much to contribute to the understanding of intelligence through general reflections on interdisciplinary issues and especially through examination of the ethical implications of developments concerning machines, animals, and people.
The Needs and Capacities of Animals Should Be Respected but not Exaggerated
Consideration of the ethical treatment of animals should avoid romantic assumptions that they are morally equivalent to humans but also shun killjoy skepticism about their capacities. The Attribution Procedure builds on substantial experimental evidence to provide good reasons why many kinds of animals are capable of consciousness, including degrees of pain, suffering, and emotion. This capability implies that animals have moral standing that should constrain their uses for food and medical experimentation. Human encroachment on natural habitats is threatening the extinction of many species which is especially tragic with the great apes whose mental capacities are close to humans in many respects.
However, the substantial mental advantages of humans show that pets are not people so that their companionship does not amount to slavery. Well-treated animals whose needs are satisfied can legitimately contribute to human happiness. There should be ongoing examination of the extent to which the limitations of animals with respect to cognition and emotion justify their uses for food and medical information. I would argue, for example, that bees do not suffer from the production of honey.
Artificial Intelligence Should Be Closely Monitored to Ensure that It Satisfies Human Needs
Intelligent machines will increasingly change human lives, but their many limitations with respect to cognition, emotion, and consciousness show that worries about whether they deserve moral standing can be postponed. Similarly, there are no immediate prospects for the development of artificial general intelligence, so we can calm fears about the prospect of a robot apocalypse that subordinates people to computers. Nevertheless, ongoing reflection on the prospects and threats of human-level AI is worthwhile. It is not too soon to identify ways of handicapping AI to prevent it from usurping humans, for example by limiting causal understanding, emotions, and creativity.
More immediately, technological developments are already posing serious threats to human wellbeing. Killer robots are increasingly feasible but should be banned internationally because they violate the ethical principles of avoiding harm to people and maintaining their autonomy. These principles also justify close monitoring of ways in which intelligent computers can make biased inferences that increase inequality. AI can also reduce freedom through violations of privacy that foster social control by governments and corporations. New applications such as driverless cars and automated decision making can thwart the productive employment that contributes to human needs for accomplishment and social connection.
My main overall recommendation is summarized in the slogan “need not greed”. Computers and robots currently do not have any needs so they do not figure in our ethical deliberations. But some animals have psychological as well as biological needs that should be taken into account. More strongly, the features and mechanisms of human intelligence generate complex psychological needs for autonomy, competence, and relatedness. Future treatment of animals and computers should be driven by considerations of biological and psychological needs, not by the greed for wealth and power that operates in selfish individuals and uncaring organizations. Compassionate people can work with bots and beasts to promote general flourishing.
This page updated July 14, 2020.