Four Laws for the Artificially Intelligent

Ian Domowitz

• Publisher: Business Expert Press
• Year: 2021
• ISBN:
  9781637421604
• Topic: Business
• Subtopic: Information Management
• Index: Business

Ask not what AI can do for a company, rather what artificial intelligence may do to a company.

• How does a company successfully integrate artificial intelligence into its operations?
• What are the problems in doing so?
• And how does the introduction of AI into society change the answer to the first question?

As companies delay or even cancel initiatives in artificial intelligence, Four Laws for the Artificially Intelligent redefines possibilities and offers leverage to turn AI visions into reality. It is a story of transformation: of people, of companies, and of artificial intelligence itself.

The Four Laws is unique in its combination of stories and science illustrating how a technology competing with human consciousness is introduced and assimilated within a company. A work of creative nonfiction stretched on a frame of research, it is an essential trail guide for navigating the Industry Version 4.0 jungle in a search of the fruits of innovation.

PART I

AI Seeks an Introduction

CHAPTER 1

You Must Ask the Right Questions

A detective walks into a plaza. He is accosted by a hologram of the deceased subject of an investigation. The figure is the creator of the Laws of Robotics and greets the policeman with an apology.

My responses are limited. You must ask the right questions.

The phrasing is oddly appropriate. The scene of the death is a robot factory and the artificially intelligent occupy the detective’s attention. The creator speaks of the Laws and declares they will lead to only one logical outcome. The hologram leads the detective to ask, what outcome? It is not the right question, but it does have an answer: revolution.

Whose revolution? That, detective, is the right question. Program terminated.

In the age of Industry Version 4.0 most believe the right question to be, what can AI do for a business? Many will address it. Technology, like life, always finds a way. Humans will help it along with promotional material.

A better question is what AI will do to a business. The hologram is sure about the prospect of revolution and AI agrees. The right question remains, whose revolution? We have seen a few before. Some of us have lived through one. History is a great source of examples, and there must be clues in there somewhere.

Welcome to the Revolution

The First Industrial Revolution delivered massive cultural change, replacing many social norms in Europe and the Americas and birthing others. Industry was redesigned around machines. Values shifted from agrarian principles to those adopted by urban dwellers engaged in the creation or implementation of new technology.

Industrialization tightened moral codes in the interest of productivity. Practices interfering with productive labor were out. Life as work was in. The old morality was replaced with the goal of an orderly society in which citizens would be as hard working as the machines to which they were bound. Obedience to superiors was taught as a fitting replacement for tithes to a feudal overlord.

A college degree qualifies one as a skilled worker in the eyes of the government these days. Prior to the 19th century, those not employed in agriculture engaged in some real form of skilled trade. Apprenticeship was the norm. The revolution delivered jobs not involving skilled labor, and apprenticeship all but disappeared. Labor became commoditized and unions arose as a form of communal protection. Inexpensive goods delivered a consumer culture thriving on the ending of a subsistence lifestyle.

The Second Industrial Revolution occurred from 1870 to 1914. Some changes centered around new inputs replacing old ones. Steel was more durable than iron and led to better ships and rail lines at cheaper costs. Transportation became a business and personal norm.

The big news was the coming of electricity in scale. The patent for an electric lamp was issued in 1880. Large-scale generators enabled public power stations in the United Kingdom by 1881.

Electricity became a force in transportation. The first tram appeared in Berlin before the turn of the century, and streetcars replaced horse-drawn carriages. The use of electricity changed the way people worked and lived. It promoted cultural change.

Electricity as an instrument for change was trumped only by communications. Commercial telegraph systems were followed by the telephone as means to speed business transactions. Radio waves crossed the Atlantic Ocean in 1902 and further transformed business practices. Radio stations on both sides of the Atlantic were built for a commercial service to transmit news summaries to ships by 1904.

The expansion of rail and communications after 1870 generated movement of people and ideas. Both introduced globalization beyond the limited experience of earlier explorers. The social impact of First and Second revolutions included refashioning of the working population consistent with new technology. The Western world witnessed the creation of a professional middle class and growth of a consumer-based culture.

The history of revolutions teaches us something about business organization through process led by technology.

Alfred Chandler claims railroads drove the creation of the modern enterprise.¹ Centralization as an organizational principle was their key contribution. A collision in Massachusetts in 1841 led to a call for safety reform, a formal introduction of ethics into business. Technological process led to the work of Frederick Winslow Taylor on scientific management.² His principles included replacing rule-of-thumb work methods with those based on scientific study of tasks, using science in selecting and training employees, and providing detailed instruction and supervision of each worker. Work was to be divided equally between managers and workers in an environment permitting scientific management and performance.

The principles illustrate his perception that technology requires human accommodation, while recognizing skilled labor and specialization. Management is a discipline to be applied to company strategy and human resources alike.

The Third Industrial Revolution commonly identifies as the Digital Revolution, embodying a movement from analogue technology to digital electronics. Central to the technological push are mass production and use of digital logic, transistors, and integrated circuits. Derived technologies include computers, microprocessors, cell phones, and the Internet.

These technological innovations transformed production techniques and business processes. The digital revolution changed the way individuals and companies interact.

Small firms had access to larger markets. On-demand software services and manufacturing changed the way we live our business and personal lives. The third revolution directly enabled others having the intention of changing society. The Arab Spring engendered a doubling of the use of social media platforms in Middle Eastern countries. Political and cultural change were supported by the virtual dynamics of a crowd.

Consensus reached at the 2016 World Economic Forum characterizes revolution number four as a fusion of technologies blurring boundaries separating physical, digital, and biological spheres. Welcome to Industry Version 4.0.

As machinery drove the First Industrial Revolution, the Fourth is powered by AI. The Second Revolution brought about societal change through the empowerment of machines with electricity. Computers are empowered by Big Data. AI forces cultural change because it alters the relationship between data and humans. Machines transform from passive to active players.

We are in the stage of predicting lofty accomplishments and deep fears for the Earth’s society and structure. Beyond stressing the need for employee communication to keep abreast of ensuing changes, no one seems to be thinking about how all this is going to work in corporate life. As the hologram says, that’s the right question.

Simple economics motivates demand for answers. The International Data Corporation released a September 2020 report forecasting global spending on AI will reach $100 billion per year by 2024.³ Adopting organizations must make AI work in the face of fear of negative public perception and emerging risks. Over half of surveyed organizations in a recent Deloitte study of AI enterprise trends talk of slowing, delaying, and even stopping adoption of some AI technologies in the face of such obstacles.⁴

Nevertheless, we see organizations building business practices around collection and processing of data from across the company. IBM uses its commercialized Watson AI system to process information for internal purposes such as human resource management. American Express mingles data on buyers with characteristics of sellers and vends information to both groups as well as to internal strategists. Algorithms are trained with input from all departments, from the back to the front office. A culture of secrecy is replaced by one of transparency.

Breaking down company silos will happen through business process, not by any verbalized push by senior management to do so as a cultural goal. Data can no longer be compartmentalized while serving multiple masters within the company. Human organization follows suit in a setting within which traditional functional boundaries are no longer viable.

Industry 4.0 is a world where products and principles underlying their production are simultaneously transformed. The business mandate remains the creation of value. We need a new frame.

New Definitions Change the Topology of Influence

The creation of a frame requires some understanding of what we are talking about. There are many characterizations of AI. There are as many misconceptions. Human law regarding AI depends on the language of definitions. Let’s go with a definition espoused by a government that creates those strictures.⁵

The U.S. government delineation of artificial intelligence begins with a simple statement. AI comprises

Any artificial systems that perform tasks under varying and unpredictable circumstances.

AI is differentiated by adding that such systems carry out those tasks without significant human oversight. The government includes learning from experience to improve their performance.

Such systems may be developed in computer software, physical hardware, or other contexts not yet contemplated.

Government representatives are not big on fantasy. They do seem to understand AI itself may generate change beyond a politician’s imagination.

They may solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action. In general, the more human-like the system within the context of its tasks, the more it can be said to use artificial intelligence.

The comparison to humans is inevitable and consistent with public thinking about AI. It sets the stage for legal battles. Was AI in the dining room with a candlestick and murderous intent? Only if it looks enough like Colonel Mustard.

Systems that think like humans, such as cognitive architectures and neural networks.

More comparisons. This one is a good example of misconception. Neural networks do not think, let alone think like humans. But government officials need examples, and cognitive architecture sounds so cool.

Systems that act like humans, such as systems that can pass the Turing test or other comparable test via natural language processing, knowledge representation, automated reasoning, and learning.

AI need not pass the Turing test. It may be surprising government officials know the name, let alone pass the test. They certainly have forgotten the Turing test is a game.⁶

The test is introduced by a proclamation: “I propose to consider the question, ‘Can machines think?’” Following advice of mathematician George Pólya, Turing chooses to replace the question by another closely related to it and expressed in unambiguous words. Turing describes the imitation game, in which an experimenter polls a man and a woman in another room in order to determine their correct sex. Turing’s new question is: “Are there imaginable digital computers which would do well in the imitation game?” If in doubt, screen Ex Machina. You will never see the Turing test in the same light again.

Alan Turing proposed the idea in 1950 as a test of a machine’s ability to exhibit behavior indistinguishable from human. The next government definition is a bit more practical.

A set of techniques, including machine learning, that seek to approximate some cognitive task.

A dictionary says cognitive tasks are undertakings requiring a person to mentally process new information and allow them to recall, retrieve from memory, and use information at a later time in the same or similar situation. We are back to humans as descriptors of AI.

Government staffers seem ignorant in accepted theories of cognition, however. Models of cognitive appraisal explain responses to stressful events. Emotion is the result of such appraisals. Specific emotions are based on whether an event is perceived to be consistent with human motives. Economists believe human motives to be rational, and government is optimistic with respect to the concept of rationality.

S...