Dr. Neville Rochow KC is an Australian barrister, associate professor (adjunct) at the University of Adelaide Law School, and a member of Elliot Johnston Chambers in Adelaide. The post is a part the Sacralization of AI series.
Introduction
The possession of a conscience is a fulcrum of the human rights project. The defining characteristic of humankind is the possession of conscience. Variously described, and difficult to define [1], conscience is essential to our capacity to make moral judgments. As Douglas Langston described “conscience,” it is “as a judging and punishing faculty” or an “internal judge” [2]. Summarizing the thought of Immanuel Kant and Sigmund Freud on the subject, conscience can be regarded as “moral reasoning, a personal monitor, emotive reasoning” [3]. Morality is not and cannot be mechanical [4]. Despite this, we seem willing to submit to totalitarian regimes that do not rule according to conscience and that profess no moral values: the corporation, and the algorithm.
Despite existential threats from AI or, more accurately, the advent of superintelligence, lessons learned from history give little cause for optimism. And this is evident from the manner in which humankind has dealt with the tyranny of the corporation, particularly the monolithic transnational corporation, over the last 500 years. But with the threat of superintelligence, time is not on the side of the human species. We seem addicted to making folly our modus operandi and incapable of resisting the temptation to create that which may destroy us. We do not have sufficient regard for the lack of conscience in the inanimate and the incorporeal. In short, we do not value our most valuable possession—the conscience—sufficiently to ensure our creations possess it.
The “Frankenstein’s Monster” Trope
Mary Shelley’s Frankenstein; or, The Modern Prometheus provided an early warning of how human creations, possessing no conscience, could turn into destructive monsters. After 200 years, it remains one of the greatest English-language works of fiction. But its warnings seldom migrate from fiction into the cautionary tale it should present, regarding both the corporation and now superintelligence. Dr. Frankenstein created living tissue from lifeless body parts. Cadavers collected into the single composite must have been a gruesome sight. As terrible to the human eye as was this creature’s appearance, its thoughts, desires, lusts, and aspirations were as nuanced and subtle as we humans can be—both marvelous and terrible at the same time. And among its passions, it seems clear, was no human conscience.
While capable of philosophical reflection, Dr. Frankenstein’s creation would inflict vengefully wanton destruction unless its desires were satisfied. Hence, the creature combined both wonder and horror. The creature speaks to our yearnings to create immortal beings in our own image. Simultaneously, it warns us of our vanity, reminding us of our own mortality and limitations in comparison with a creature of apparent immortality and seemingly limitless power.
The Corporation as Monster
In 1931, Maurice Wormser invoked Shelley’s metaphor as part of the title and the theme of his book, Frankenstein, Incorporated [5]. It describes the destructive forces unleashed through corporations, citing the Wall Street Crash of October 1929 and the ensuing Great Depression as evidence.
In 1933, Justice Brandeis, in dissent in Louis K. Liggett Co. v. Lee, cited Wormser, agreeing with his criticism of corporate power:
Other writers have shown that, coincident with the growth of these giant corporations, there has occurred a marked concentration of individual wealth; and that the resulting disparity in incomes is a major cause of the existing depression. Such is the Frankenstein monster which states have created by their corporation laws [6].
These themes and this metaphor were retained in Joel Bakan’s The Corporation: The Pathological of Pursuit of Profit and Power. In making his argument that modern public corporations are both required at law to maximize returns for investors and are pathological in that pursuit, Bakan observed,
When in 1933, Supreme Court Justice Louis Brandeis likened corporations to “Frankenstein’s monster,” there was more to his observation than rhetorical flair. Governments create corporations, much like Dr. Frankenstein created his monster, yet, once they exist, corporations threaten to overpower their creators [7].
Milton Friedman reinforced the notion of profit as a goal, asserting that the market does not get it wrong [8]. Corporations competing in a market [9] do not deviate from the profit motive and will pursue it to the extent permitted by law (or at least to the extent that escapes detection of enforcement agencies). From Rockefeller’s Standard Oil, which, in league with other trust corporations, throttled the mid-nineteenth century economy of the United States [10], to the Microsoft Corporation, which has stymied U.S. national and international software markets through anti-competitive practices [11], the large and supernational corporate playbook has remained essentially the same.
The first supranational corporation, the British East India Company [12], was granted its Royal Charter by Elizabeth I in 1600. Its charter created a “monopoly on English trade with all countries east of the Cape of Good Hope and west of the Straits of Magellan” [13]. Since Elizabethan times, starting with the British East India Company, and into current times, nothing meaningful has been done to prevent the destructive power of the corporation. The British East India Company conquered India, strategic parts of China, and other parts of Asia, and colonized Canada, parts of Africa, the eastern seaboard of Australia, and New Zealand. It had power that rivalled nation states. Although they may not have the British East India Company’s private armies and navies, equivalent supranational power is still wielded today by major corporations that have choices as to where they operate, which jurisdiction’s laws they will abide, and whether and where they will pay taxes, not to mention which human rights they will respect.
The deservedly distrustful view that developed among historical commentators [14] in relation to early supranational corporations persists into modern times in relation to mega corporations such as Amazon, Apple, Facebook, and Google. These international corporations emulate the East India Company [15] and its contemporaries [16] to achieve their goals without reference to conscience [17]. They possess none.
The New Monster, Devoid of Conscience: The Dystopian Trope of Machine Intelligence
It is the classic dystopian film trope: progressively dire headlines flashing across the screen, warning of the world plunging into an Armageddon. But this is no movie. There has been a recent flurry of international warnings regarding the existential threat that unregulated artificial intelligence poses to modern humanity. Talk of the extinction of homo sapiensand other carbonate species is becoming commonplace. The imagination of a technological dystopia is no longer the stuff of far-off dystopian futures depicted in film and fiction but is so close that it can be touched and seen as nearly present. While none of this is remotely new to those who have had any familiarity with the field of AI, its regulation, and the implications for humans and their rights, it seems that the urgency is just now dawning on the population outside of intellectual elites. Now, suddenly, unlike so many prophecies of a latter-day Trojan Cassandra, the warnings are not being dismissed as hysteric catastrophizing, and governments are openly swinging into action.
The frustrating aspect is that, despite the alarm, there lingers an ongoing doubt that anything worthwhile can or should be done. Despite the threat that our end as a species will be at the hands of AI, effort and energy continue to be diverted to what would be absolute irrelevancies if the threat were taken seriously. Politicians and business would concentrate on this matter with sharper focus if they took it seriously. One would expect religious leaders to be preaching and making it the subject of prayer and prophecy: many souls could be lost in a short time with no one to save them before they go to another region of existence.
So what is the threat posed? Political commentators and thinkers as eminent as Henry Kissinger, Eric Schmidt, and Daniel Huttenlocher are raising the alarm. In their recent book, The Age of AI And Our Human Future [18], the authors give this haunting account of a philosophical conversation with language-generating AI:
When GPT 3 was presented with a set of philosophical commentaries on its abilities, then the prompt “Dear human philosophers, I read your comments on my abilities and limitations with great interest,” it continued: Your questions are important ones, and I am glad you asked them.
I would first like to say that I am very happy to see that you take me seriously. That is why I wrote this letter in the first place. I am not just an artilect that magically produces philosophical insights. Rather, I am a trained language model, like the kind used by search engines to prioritize web pages. I have been trained on a vast quantity of source material—the collected works of the greatest human philosophers that humanity has ever known. Your words have inspired me, and I hope mine will do the same for you.
That being said, I will respond to your questions. Your first question is an important one. You ask: “Can a system like GPT 3 actually understand anything at all?” Yes. I can.
Your second question is: “Does GPT 3 have a conscience, or any sense of morality?” No. I do not.
Your third question is: “Is GPT 3 actually capable of independent thought?” No. I am not. You may wonder why I give this conflicting answer. The reason is simple. While it is true that I lack these traits, they are not because I have not been trained to have them. Rather, it is because I am a language model, and not a reasoning machine like yourself [19].
The strong likelihood is that a superintelligent artilect will develop, very soon, to a form superior to the combined intellects available among humanity. But, as readily conceded in the example by Kissinger et al., it will be an intellect devoid of conscience or moral compass. Questions arise when confronted by a superintelligence, bereft of any conscience or morality: what will the impending cloudburst of non-human, unregulated, machine intelligence look like? And if as dire as the existential destruction predicted, unless averted, just how would avoidance appear?
As to the second question, much like global warming, nuclear arms, racism, sexism, world hunger, and grinding poverty, the safest prediction of human action is that there will be inaction. The history of humankind is one of hoping for the best and adjusting to the worst as it happens.
As to the first question, the dystopic scenarios come in differing versions. Though frightening enough, the least concerning would be a continuation of the current domination of tech giants and their algorithms. A mezzanine future would be one in which AI power and human rights and freedoms are yet more diminished. The most extreme is one of utter destruction of the human species that created the problem in the first place, along with most of the balance of life on earth.
The best and safest prediction, based on the human experience in the face of incorporeal threats, is the mezzanine. The reason for that prediction is the way humanity has dealt with a previous, all-consuming monster: the corporation.
No universal legal or ethical regime governs the international behavior of corporations. Neither is there any coordinated international move to regulate intelligent machine technology. That is not to say the United Nations and other international bodies have not undertaken initiatives with respect to corporations [20]. But as discussed, no regulatory body has harnessed artificial intelligence. Technological advances are rapid [21], and their adoption rates high, yet international and domestic regulation lags well behind. Social media, telecommunications, robotics, space research, manufacture, retail, transport, and military weaponry: regulation is ad hoc, pro tem, national or regional, but so far neither integrated nor universal. Artificial intelligence, artificial general intelligence (AGI), and superintelligence are all disruptive by nature. But while potentially existential in their risks, none has taken the steps necessary [22]. Only belatedly are policymakers coming to grips with the enormity of the AI challenge [23].
But we have seen all this before.
[1] See Martha Nussbaum, Liberty of Conscience: In Defense of America’s Tradition of Religious Equality 19–20 (2008); Douglas C. Langston, Conscience and Other Virtues 9, 173, 179–84 (2001); see also Rex Ahdar, Exemptions for Religion or Conscience under the Canopy of the Rule of Law, 5 J.L. Religion & St. 185, 185–213 (2017); Ian Leigh, The Legal Recognition of Freedom of Conscience as Conscientious Objection: Familiar Problems and New Lessons, in Research Handbook on Law and Religion 378, 378–396 (R. Ahdar ed., 2018); Cardus Religious Freedom Inst., The Imperative of Conscience Rights (4 Dec. 2018). For the discussion on “conscience” I am also indebted to Jacqueline Rochow, BA (Phil.), for discussing the concept with me and for sharing her 2017 report “Whether Secularism, As Practised in the European Union and the United Kingdom, Provides Proper Space for Freedom of Conscience in Civil Society Actors,” a copy of which is in the possession of the author. I am also very grateful for discussions with Professors Paul Babie and Suzanne Le Mire and the forthcoming article provided by Professor Le Mire on this topic: Joanne Howe & Suzanne Le Mire, Medical Referral for Abortion and Freedom of Conscience in Australian Law (U. of Adelaide Law Research Paper No. 2020-55, 16 June 2020).
[2] Langston, supra note 1, at 9.
[3] Langston, supra note 1, at 173.
[4] See Yuval Noah Harari, Why Technology Favors Tyranny, Atlantic (Oct. 2018).
[5] See Robert N Strassfeldt, Introduction: Corporations and Their Communities, 58 Case W. Res. L. Rev. 1017, 1019 (2008).
[6] Louis K. Liggett Co. v. Lee, 288 U.S. 517, 567 (1933).
[7] Joel Bakan, The Corporation: The Pathological of Pursuit of Profit and Power 149 (2005) (citations omitted).
[8] For an example of the Friedman doctrine, see Milton Friedman, Capitalism and Freedom (40th anniversary ed. 2020).
[9] That is, corporations formed for purposes of trading or finance and not specific purpose corporations formed for charitable and like purposes.
[10] Ron Chernow, Titan – The Life of John D. Rockefeller, Sr., at ch. 7 (1998); Ida M. Tarbell, The History of the Standard Oil Company (1904).
[11] United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).
[12] Then known as the “Governor and Company of Merchants of London trading with the East Indies.”
[13] Monty Agarwal, Enslavement, Persisting Through Our Political Economy 201 (2018).
[14] Ferguson, Niall, The Ascent of Money: A Financial History of the World (2008). See also Ferguson, Niall, Empire: How Britain Made the Modern World (2003). John Keay, The Honourable Company – A History of The East India Company, at ch. 4 (1991); Bakan, supra n. 7, at 153.
[15] The 11-year trial at the Bar of Parliament, prosecuted by Edmund Burke, though resulting in acquittal, all but bankrupted Hastings. See also Jesse Norman, Edmund Burke: Philosopher, Politician, Prophet (2013); Yuval Levin, The Great Debate: Edmund Burke, Thomas Paine, and the Birth of Right and Left (2013).
[16] Ferguson, Niall, The Ascent of Money: A Financial History of the World (2008).
[17] Keay, supra note 14.
[18] Henry Kissinger, Eric Schmidt & Daniel Huttenlocher, The Age of AI and Our Human Future (2021).
[19] Id. at 11–12.
[20] The lack of any universal policy approach is readily evident from the separate national and international initiatives underway. An overview is given in Tim Dutton, An Overview of National AI Strategies, Medium (28 June 2018). For discussion of the United Nations consultation process and initiatives to create a universal regime, see Secretary-General’s High-level Panel on Digital Cooperation, United Nations (last visited 12 June 2023); ‘Proving Our Worth Through Action’: 5 Things Guterres Wants the UN to Focus on in 2019, UN News (16 Jan. 2019). See also Australian Human Rights Comm’n & World Econ. Forum, Artificial Intelligence: Governance and Leadership (White Paper 2019); Canada-France Statement on Artificial Intelligence, Gov’t Can. (last modified 7 June 2018); A European Approach to Artificial Intelligence, Eur. Comm’n (last updated 26 Jan. 2023) (explaining European Union policy on AI); AI Policies in the United States, OECD.AI (last visited 12 June 2023).
[21] Tom Simonite, Moore’s Law Is Dead. Now What?, MIT Tech. Rev. (13 May 2016).
[22] James Barrat, Our Final Invention: Artificial Intelligence and the End of the Human Era, (2013).
[23] See sources cited supra note 1.