Unjust Law is Itself a Species of Violence: Oversight vs. Regulating AI
Summary
The cases and ideas discussed, from Louis D. Brandeis's advocacy for transparency to Charles Lindbergh's belief in the marvels of exploration and Alan M. Turing's pioneering work in computer science, collectively emphasize the merits of oversight and transparency over heavy regulation in emerging technologies such as artificial intelligence. They underscore the importance of informed decision-making and data-driven governance in guiding innovation without stifling it. Striking this balance allows technology to venture into uncharted territory, promising progress and marvels while ensuring responsible development and safeguarding against potential misuse. Innovation and progress often require a degree of freedom from strict regulation or confinement. Justice Louis Brandeis one said "Behind every argument is someone's ignorance."
Challenging the status quo and pushing boundaries often stem from questioning prevailing norms and regulations. Innovation thrives when individuals are free to challenge existing constraints. Generative AI resides firmly within this realm, offering a glimpse of what lies on the horizon as information technology advances at an exponential pace. With each passing day, we uncover new capabilities within Foundation Models, and yet, this is merely the inaugural chapter in the story of AI's evolution
Penetrating the Unknown
“In some future incarnation from our life stream, we may even understand the reason for our existence in forms of earthly life. The growing knowledge of science does not refute man's intuition of the mystical. Whether outwardly or inwardly, whether in space or in time, the farther we penetrate the unknown, the vaster and more marvelous it becomes.”
Charles Lindbergh's profound insight into the mystical and the unknown can be thought of as a powerful metaphor for the realm of artificial intelligence (AI) and its potential. AI, a field that emerged in earnest only in the mid-20th century, particularly with the groundbreaking work of Alan Turing, represents a relatively recent addition to the vast landscape of human knowledge and innovation. Lindbergh's words remind us that the deeper we delve into the unknown, whether it's in space, time, or new scientific disciplines, the more vast and marvelous it becomes.
Law of Accelerated Returns & Information
Technology makes us more human, not less. Technology amplifies are consciousness if used in the right direction. Law of accelerated returns existed even before humans invented wheels since before humans invented wheels they invented a way to communicate - that’s information. Law of accelerated returns applies to anything information. So as long as we continue to use the information in the right direction, technology will continue to amplify our abilities and make us more human. “a person on a bicycle was more energy efficient than a condor in flight and many times more energy efficient than a person in an automobile”
Quoting Steve Jobs “I read a study that measured the efficiency of locomotion for various species on the planet. The condor used the least energy to move a kilometer. And, humans came in with a rather unimpressive showing, about a third of the way down the list. It was not too proud a showing for the crown of creation. So, that didn’t look so good. But, then somebody at Scientific American had the insight to test the efficiency of locomotion for a man on a bicycle. And, a man on a bicycle, a human on a bicycle, blew the condor away, completely off the top of the charts. And that’s what a computer is to me. What a computer is to me is it’s the most remarkable tool that we’ve ever come up with, and it’s the equivalent of a bicycle for our minds.”
The “somebody” at Scientific American was S. S. Wilson and the eleven-page article in question, on bicycle technology, was printed in the March 1973 edition of the magazine. Wilson was a lecturer in engineering at Oxford University and a fellow of St. Cross College. S. S. Wilson said: “My interest in bicycles dates back to school days. I have always owned and used a bicycle; during World War II, I several times cycled more than 100 miles in a day as a means of transport.”
Information and its related law of accelerated returns have been part of human civilization for 1000s of years and common law doctrine in the USA or UK and Napoleonic Code (Civil Code of the French in 1804) or a variation of it in regions such as France, India and generally Asia, takes nuances of information and hence information technologies into consideration. Birth of computer science in mid-20th century make it information technology scientific and to some extent more repeatable.
Computer Science
The birth and conception of computer science, as we know it today, owe a great deal to the pioneering work of Alan M. Turing, often hailed as the father of computer science. Alan Turing's seminal contributions, particularly his concept of the Turing machine, laid the foundation for both theoretical and practical aspects of computing. In the mid-20th century, during a time when the world was grappling with the challenges of World War II, Turing's innovative thinking led to the development of the Bombe machine, used to decipher German Enigma machine-encrypted messages, a crucial contribution to the Allied victory.
Turing's genius extended far beyond wartime cryptography. His work on the Universal Turing Machine, introduced in his 1936 paper "On Computable Numbers," provided the theoretical framework for modern computers. The Turing machine was a conceptual model of computation, capable of performing any mathematical algorithm, making it a precursor to today's general-purpose computers. Turing's ideas revolutionized our understanding of computation and set the stage for the digital age and the discipline that has emerged recently in the last 20 years, artificial intelligence (AI)
AI & Regulation
Regulating AI today, while well-intentioned, has the potential to hinder innovation in this nascent field. Much like the early days of computer science when Turing laid the foundation for AI, the field is still evolving and exploring its boundaries. Overregulation can stifle creativity, experimentation, and the exploration of uncharted territories within AI. It risks locking AI development into rigid frameworks that may not accommodate future breakthroughs and discoveries.
The essence of AI lies in pushing the boundaries of what's possible, in developing algorithms that can process vast amounts of data, in creating machines that can learn and adapt, and in solving complex problems that were once considered insurmountable. These pursuits are inherently speculative, and the outcomes often exceed our initial expectations.
By imposing strict regulations prematurely, we risk inhibiting the very innovation that has the potential to unlock the next marvels of AI. We may miss opportunities to develop technologies that can improve healthcare, enhance education, address environmental challenges, and revolutionize industries. AI's ability to penetrate the unknown and tackle previously unsolvable problems is one of its greatest promises.
“The Marketplace of Ideas" Principle
"The marketplace of ideas," a concept closely associated with the renowned philosopher John Stuart Mill, holds that addressing harmful or false notions should not involve censorship or regulation. Instead, it champions letting these ideas contend in an unfettered exchange of thoughts and opinions within society.
According to Mill's perspective, as expounded in his influential essay "On Liberty," published in 1859, this open discourse allows good ideas to naturally triumph over bad ones. Mill's ethical framework of utilitarianism, which underpins his arguments in "On Liberty," extends the principle of individual freedom to society and the state. He believed that by permitting diverse and even contentious ideas to circulate freely, societies would ultimately benefit from the intellectual competition, fostering progress, and the betterment of society. This approach to liberty and open debate remains foundational in discussions of free speech and the regulation of ideas in modern democracies. As Mill eloquently stated in his work,
"The worth of a state, in the long run, is the worth of the individuals composing it."
This is also the right way to think about the recent development in generative AI and more generally AI. A state neither has the expertise nor the means to censor or regulate these technologies and related ideas. Instead the role of the state should be to facilitate and create the right environment for the unfettered exchange of such ideas, which is essential to the preservation of a free and democratic society.
Oversight
Oversight and regulation are related. Oversight involves monitoring and supervision, often carried out by a governing body or authority, to ensure that existing regulations and laws are being followed. Regulation, on the other hand, encompasses the creation and enforcement of rules, laws, and policies.
Northern Securities Co. v. United States, a U.S. Supreme Court case in 1904, left an indelible mark on American antitrust law. In a closely divided 5-4 decision, the Court ruled against the stockholders of the Great Northern and Northern Pacific railroad companies, who had effectively created a monopoly through the formation of the Northern Securities Company. This landmark case exemplified the U.S. government's commitment to curbing monopolistic practices that threatened healthy competition within the marketplace. The Court's decision to dissolve the Northern Securities Company signaled a significant victory in the ongoing battle to preserve economic competition and consumer choice, setting a precedent for antitrust enforcement that would shape the course of American business regulation in the years to come.
Supreme Court Justice, Oliver Wendell Holmes, Jr. wrote in his dissenting opinion in the case of Northern Securities Co. v. United States (1904)
"I recognize without hesitation that judges do and must legislate, but they can do so only interstitially; they are confined from molar to molecular motions,"
underscores the idea that judges can interpret and apply existing laws (oversight) to address specific issues without the need for new legislative regulation. In this context, the U.S. Supreme Court's decision in the Northern Securities case served as an oversight mechanism, with the Court interpreting and applying the Sherman Antitrust Act to dissolve the monopoly, thereby upholding the principles of competition and preventing the need for new, specific regulations targeting this particular case.
This historical example illustrates how a combination of existing laws, oversight, and judicial interpretation can effectively address anticompetitive practices and promote transparency within the business environment. It demonstrates that a well-established legal framework, when rigorously enforced and applied, can serve as a robust tool for maintaining fair competition and preventing the undue concentration of economic power without the necessity of continually enacting new regulations.
Regulation
In Northern Securities Co. v. United States, the U.S. government applied the Sherman Antitrust Act of 1890 to scrutinize the actions of the Northern Securities Company, which had effectively created a monopoly through the consolidation of two major railroad companies.
John Sherman, the American politician and lawyer responsible for the Sherman Antitrust Act of 1890, stated:
"If we will not endure a king as a political power, we should not endure a king over the production, transportation, and sale of any of the necessaries of life."
Sherman's quote reflects the intent of the Sherman Antitrust Act, which aimed to regulate and prevent monopolistic practices in commerce. It demonstrates the concept of regulation to prevent anti-competitive behavior.
Some of the current regulations that are implemented in the AI regulations space:
General Data Protection Regulation (GDPR): While not specific to AI, GDPR, implemented in the European Union in 2018, includes provisions related to automated decision-making and profiling, which can be relevant to AI systems that process personal data.
Ethical Guidelines for Trustworthy AI by the European Commission: In 2019, the European Commission published guidelines outlining seven key requirements for trustworthy AI, including transparency, accountability, and fairness. These guidelines are non-binding but provide a reference for the ethical development and deployment of AI systems.
Algorithmic Accountability Act (AAA): Proposed in the United States, the AAA aims to address potential biases and discriminatory impacts of automated decision-making systems. It seeks to require companies to assess and mitigate biases and provide transparency in their algorithms.
Algorithmic Transparency and Accountability in Public Administration and Services (ATAP): Proposed by the European Parliament, ATAP aims to ensure transparency, accountability, and explainability of AI systems used in public administration and services.
AI Act: In April 2021, the European Commission proposed a comprehensive regulatory framework for AI, known as the AI Act. It aims to establish legal requirements for AI systems' development, deployment, and use within the European Union, focusing on high-risk AI applications.
Montreal Declaration for Responsible AI: The Montreal Declaration, introduced in 2018, is an international initiative promoting the development of AI that respects human rights, diversity, and environmental sustainability.
National AI Strategies: Various countries, including Canada, China, France, Germany, and the United States, have developed or announced national strategies for AI. These strategies typically encompass a range of aspects, including research funding, workforce development, and policy frameworks.
Transparency
Louis D. Brandeis, an influential U.S. Supreme Court Justice in the early 20th century, was renowned for his advocacy of transparency and consumer protection. His jurisprudence left a lasting impact on American law, particularly in the realms of privacy and antitrust regulation.
During his time as a Supreme Court Justice, he was a strong proponent of transparency and the right to privacy. One of his most notable writings on transparency is his 1914 book "Other People's Money and How the Bankers Use It." In the book he wrote
"Sunlight is said to be the best of disinfectants; electric light the most efficient policeman."
In this work, Brandeis argued for greater transparency and oversight in the financial industry, particularly in the regulation of investment and banking practices. Brandeis was an advocate for oversight and transparency as means to curb the excesses of powerful financial interests. His quote suggests that oversight (in the form of transparency) can be a powerful tool to prevent wrongdoing. His opinions and writings, including the famous "Brandeis Brief" in the case of Muller v. Oregon (1908), emphasized the importance of social and economic data as a basis for legal decisions. In Muller v. Oregon, he utilized sociological data to support the state's regulation of working hours for women, highlighting how transparency through data and research could inform legal decisions and promote social welfare.
Brandeis also authored the influential dissenting opinion in the case of Olmstead v. United States (1928), which argued for the protection of privacy rights, particularly in the context of wiretapping and surveillance. Although a dissenting opinion, his views laid the groundwork for future legal developments related to privacy and transparency in the digital age. Louis D. Brandeis's writings and legal philosophy significantly contributed to the broader discussion of transparency, privacy, and the role of government in regulating business practices during his time on the Supreme Court.
Oversight & Transparency Instead of Regulation
Responsible oversight and ethical considerations are crucial in AI development to prevent misuse and protect individuals' rights. Given that the field of computer science itself is relatively young, dating back to the mid-20th century, it underscores the notion that regulating AI, a subset of this discipline, too aggressively at this stage could hinder the spirit of exploration and innovation that characterized its early days. Alan Turing's legacy reminds us of the transformative power of groundbreaking ideas and the need to foster an environment that encourages creative thinking and the pursuit of the unknown in AI and computer science.
Issues in the past have originated from lack of oversight and not absence of a regulation. This lack of oversight has resulted in outcome has been Google and Facebook with M&A that has stiffened competition and startup development. EU is a great example which today has been in forefront of regulations but it lagging behind in startup ecosystem for decades and that continues to be the case today.
In the US, both OpenAI and Google announced last week on how website administrators can prevent their bots from scraping individual websites for AI training purposes. OpenAI has uses web crawler (GPT Bot) and last week provided guidance to website publishers on how to prevent the bot from scraping their websites. One effective method for website owners to protect their content is by using a robots.txt file to disallow access to specific parts or the entirety of their site. However, additional measures like CAPTCHAs, traffic pattern monitoring, and IP blocking may be required for more robust protection against unauthorized data harvesting. Additionally, Google has introduced a tool called Google-Extended, allowing websites to be indexed by crawlers while keeping their data inaccessible for training future AI models, offering website administrators greater control over their content's use for generative AI applications. This was not done by OpenAI and Google due to an AI specific law that exist in the US, but because of making their product better, competitive pressures, consumer sentiments as well as oversight by both regulators and concerned citizens.
AI and data privacy specific lawsuits against Google:
J.L. v. Alphabet Inc. (filed July 11, 2023): Class-action lawsuit alleging that Google is misusing personal data and copyrighted material to train its AI products without consent.
Clarkson Law Firm v. Google LLC, et al. (filed July 11, 2023): Class-action lawsuit alleging that Google's privacy policies are misleading and violate users' privacy rights.
In addition, general US lawsuits against Google:
Texas v. Google (filed December 2021): Antitrust lawsuit alleging that Google has abused its dominance in the search and app markets.
Utah v. Google (filed December 2021): Antitrust lawsuit alleging that Google has abused its dominance in the app market.
States of America v. Google (filed January 2023): Antitrust lawsuit alleging that Google has abused its dominance in the online advertising market.
Justice Department v. Google (filed January 2023): Antitrust lawsuit alleging that Google has abused its dominance in the online advertising market,
and outside the US lawsuits:
European Commission v. Google (ongoing): Antitrust investigation into Google's use of its search engine to promote its own products and services.
French Competition Authority v. Google (ongoing): Antitrust investigation into Google's use of its search engine to promote its own products and services.
German Federal Cartel Office v. Google (ongoing): Antitrust investigation into Google's use of its search engine to promote its own products and services.
UK Competition and Markets Authority v. Google (ongoing): Antitrust investigation into Google's use of its search engine to promote its own products and services.
Lawsuits against OpenAI related to AI, privacy, and data.
US lawsuits against OpenAI:
PM v. OpenAI LP (filed August 4, 2023): Class-action lawsuit alleging that OpenAI violated privacy laws by collecting and using users' personal data without consent to train its AI models.
Clarkson Law Firm v. OpenAI LP (filed June 20, 2023): Class-action lawsuit alleging that OpenAI's web scraper violated privacy rights and property rights of individuals whose personal information was scraped and then incorporated into OpenAI's products.
Sarah Silverman et al. v. ChatGPT, et al. (filed July 18, 2023): Lawsuit alleging that ChatGPT maker OpenAI and Meta are using copyrighted work without permission in AI training.
The New York Times v. OpenAI (pending): Lawsuit alleging that OpenAI is infringing on The New York Times' copyright by using its news articles to train its AI models.
Outside the US lawsuits against OpenAI:
OpenAI and Microsoft hit with new US consumer privacy class action (September 6, 2023): Lawsuit alleging that OpenAI and Microsoft violated privacy laws in developing OpenAI's popular chatbot ChatGPT and other generative artificial intelligence systems.
The historical cases and principles advocated by figures such as Louis D. Brandeis, Charles Lindbergh, and Alan M. Turing underscore the value of oversight and transparency over heavy-handed regulation, particularly in the realm of technologies that venture into the uncharted territories of innovation. Brandeis's advocacy for transparency and informed decision-making in cases like Muller v. Oregon resonates with the idea that thorough oversight, coupled with the collection and analysis of relevant data, can guide effective regulation without stifling progress. In this context, the wise words of aviator Charles Lindbergh remind us that as we delve deeper into the unknown, the potential for marvels and progress expands. This notion applies aptly to the evolving field of technology, where innovation should flourish under the guiding principles of oversight and transparency rather than excessive regulation.
Similarly, Alan M. Turing's groundbreaking work in computer science exemplifies the monumental strides that can be achieved when innovation is allowed to flourish within a framework of oversight, ethics and transparency. By maintaining this balance, we empower AI to penetrate the unknowns past Generative AI constructs such as LLMs (Large Language Models) based on the Transformer architecture, promising a future where the marvels of AI innovation continue to enrich society while safeguarding against misuse and harm.
Conclusion
In conclusion, our exploration of the concepts of oversight, transparency, and regulation in the context of emerging technologies, such as Generative AI, reveals a delicate balance that holds the key to fostering innovation while safeguarding against potential risks. From the wisdom of legal minds like Louis D. Brandeis, who championed transparency as a means to informed decision-making, to the visionary spirit of pioneers like Charles Lindbergh, who celebrated the marvels of venturing into the unknown, we find a common thread. This thread extends to the realm of technology, as exemplified by the groundbreaking work of Alan M. Turing, where innovation flourishes when nurtured within a framework of oversight and ethical considerations.
As we chart the course for generative AI and other emerging fields to follow, the resounding message is clear: it is through the power of oversight and transparency that we can harness the full potential of the unknown, unlocking a future where progress and marvels coexist with responsible development.
Mahatma Gandhi in his book Non-Violence in Peace & War (1942), said underscoring the unjust nature of some laws and the violence inherent in enforcing them.
“An Unjust law is itself a species of violence. Arrest for it's breach is more so. Now the law of non-violence says that violence should be resisted not by counter-violence but by non-violence. This I do by breaking the law and by peacefully submitting to arrest and imprisonment.”
Further read
Applying Asimov's Laws to Robot Taxis Confronted by Threats.
From Infinite Improbability to Generative AI: Navigating Imagination in Fiction and Technology
Human vs. AI in Reinforcement Learning through Human Feedback.
Generative AI for Law: The Agile Legal Business Model for Law Firms
Generative AI for Law: From Harvard Law School to the Modern JD