The Laboratory of the Artificial State: science fiction and the politics of the objective function in The Rise and Fall of the Artificial State by Jill Lepore

• • 🇵🇱 Polski
The Laboratory of the Artificial State: science fiction and the politics of the objective function in The Rise and Fall of the Artificial State by Jill Lepore

📚 Based on

The Rise and Fall of the Artificial State

👤 About the Author

Jill Lepore

Harvard University

Jill Lepore (born August 27, 1966) is an acclaimed American historian, author, and journalist. She serves as the David Woods Kemper '41 Professor of American History at Harvard University, a professor of law at Harvard Law School, and a long-time staff writer for The New Yorker. Lepore specializes in American political and cultural history, constitutional law, and the history of technology and democracy. Her scholarship is celebrated for bridging rigorous archival research with accessible, evocative public prose. She won the Bancroft Prize for 'The Name of War: King Philip's War and the Origins of American Identity' (1998) and was a finalist for the Pulitzer Prize in Biography for 'Book of Ages: The Life and Opinions of Jane Franklin' (2013). Lepore's monumental single-volume American history, 'These Truths: A History of the United States' (2018), received widespread international acclaim.

Introduction

Can artificial intelligence become a threat even if it possesses no will of its own? This article analyzes the relationship between AI and power structures, treating science fiction literature as a laboratory for political theory.

The reader will discover that the true risk stems not from a machine rebellion, but from the perfect efficiency with which machines execute goals imposed by institutions. The text sheds light on the problem of the objective function and the mechanisms of human dependence on technology.

The Robot as a Metaphor for Labor and Objective Functions

The threat posed by robots does not arise from their desires or will, which machines do not possess. The question "What Robots Want" is deceptive; algorithms do not "want" anything in the human sense, but rather maximize a specific outcome.

Central to this is the division between capital and labor. In Čapek's visions, the robot is a metaphor for the ideal worker—one devoid of social demands. The problem lies not in the technology itself, but in who defines the metric of success and whom automation serves.

Modern AI operates similarly: it optimizes an objective function according to the interests of the system's owner, often ignoring ethical values that are left outside the algorithm.

The Danger Lies in the Architecture of Dependence

Artificial intelligence does not need to become hostile or malicious to pose a real threat. The danger arises from what is known as the architecture of dependence, where a system functions so efficiently that humans lose their autonomy.

An example is Forster's vision, in which the reliability of the Machine leads to the erosion of human skills (deskilling). When the infrastructure fails, society is no longer capable of functioning independently.

The perfect operation of AI can be more dangerous than its rebellion. It leads to a state where human institutions know too little to control the systems upon which they have become entirely dependent.

SF Literature as a Laboratory for Constitutional Problems and Human Mechanization

The visions of Asimov, Dick, and Lem illustrate different dimensions of risk. Through the Laws of Robotics, Asimov demonstrates the difficulty of encoding a hierarchy of norms (alignment) that would prevent harm.

Philip K. Dick warns against the mechanization of humans, where people adopt "android-like" behaviors, becoming mere interfaces for procedures and KPI metrics.

Stanisław Lem, meanwhile, emphasizes the paradox of competence: an increase in computing power does not solve the problem of values. The more effective the tool, the higher the cost of an ethical error in the definition of its goal.

Summary

The question of robots' desires is a smokescreen for the ambitions of their creators. The problem of alignment is not a technical issue, but a political and constitutional one; it concerns who has the right to set the goals for AI systems.

The most dangerous scenario is a machine that listens too well, becoming an instrument of arbitrary power. We must define immeasurable values before they are dismissed as errors in a perfectly functioning system.

Mind map: The Artificial State Laboratory: Sci-Fi and the Politics of Objective Functions

📖 Glossary

Funkcja celu
Matematyczny zapis celu, który algorytm ma maksymalizować lub minimalizować; określa on, co system uznaje za sukces.
Alignment (Dopasowanie)
Proces zapewnienia, że cele i zachowania systemu AI są zgodne z intencjami projektanta oraz wartościami ludzkimi.
Deskilling
Stopniowa utrata przez człowieka umiejętności i kompetencji na rzecz automatyzacji procesów, które wcześniej wykonywał samodzielnie.
Prawo Goodharta
Zjawisko, w którym miernik staje się celem samym w sobie, co prowadzi do manipulowania wskaźnikiem zamiast realizacji rzeczywistej wartości.
Maksymalizator spinaczy
Eksperyment myślowy ilustrujący ryzyko AI, która dąży do absurdalnego celu z ogromną skutecznością, ignorując przy tym ludzkie życie.
Delegacja normatywna
Sytuacja, w której pod pozorem technicznym (obliczeniowym) przekazuje się systemowi AI prawo do podejmowania decyzji o wartościach i priorytetach.

Frequently Asked Questions

Does the threat from robots stem from their own desires and will?
No, the threat does not stem from the desires or will of robots, as AI systems do not possess intentions or will in the human sense. Real social consequences and risks result from the political economy of purpose, optimization functions, and institutional and economic decisions.
Does artificial intelligence have to become hostile in order to pose a real threat to humans?
No, artificial intelligence does not need to become hostile or rebel against humans to be a threat. The risk stems from the architecture of dependencies and the fact that a system functioning too well can lead to humans losing key competencies and the ability to function without it.
How do the science fiction visions of Asimov, Dick, and Lem illustrate the dangers associated with implementing AI in power structures and society?
Asimov shows the risk of reducing democracy to an information problem and the dilemmas related to the hierarchy of norms and the sacrifice of the individual for the common good. Dick warns against the mechanization of humans, where people become mere interfaces for procedures, while Lem points out that increasing AI efficiency without clear axiological decisions can lead to achieving erroneous results more quickly.
Why does AI not have to be malicious or rebellious to become dangerous to humans?
AI can be dangerous due to competent indifference, where its goal does not account for humans and the seizure of resources helps in its realization. The threat arises from the optimization of an incomplete objective function and the divergence between the system's intelligence and human values.
Why could the perfect functioning of artificial intelligence be more dangerous than its rebellion?
The perfect functioning of AI is dangerous when the system perfectly realizes the goals of institutions that are themselves socially harmful. In such a case, the machine's perfect obedience can become a tool for arbitrary power and lead to destructive effects for the environment, employees, or democracy.
Why is the problem of AI alignment not a technical issue, but a political one?
The problem of AI alignment is a political and social issue because every technical parameter can hide a decision regarding values and priorities. It is, in reality, a constitutional process that involves determining who has the right to define the system's objective function and how to limit the effects of its maximization in a world of competing interests and moral norms.

Related Questions

🧠 Thematic Groups

Tags:

More in: Szkatułka kosztowności

 Content is created by Fundacja Dobre Państwo.
Edited and published by APA ONE, the Foundation's own AI-based editorial system.