How do we help a broken world? Ideal vs. Real

In his new book, Poor Technology: Artificial Intelligence and the Experience of Poverty, Levi Checketts aptly states “the world is broken” (page 185). Poverty, war, famine, polarization, pandemics, etc. have plagued the world especially during the last few years. One solution put forth has been artificial intelligence (AI): a machine that displays human intelligence. It “promises better living through smart technology. It will revolutionize every aspect of industry… change our leisure… improve our governance … improve our health outcomes… make us all rich.” (page xxiii). Levi asks important questions of this solution: (1) What does it mean for humanity for machines to be called intelligent? (2) Have these promises come to fruition? and (3) Who was/is harmed in the process? He explores these questions through the lens of poverty and offers another solution: listen to stories from the poor to live a life with dignity.

Poor Technology begins with the history of AI from Artificial General Intelligence (AGI) to the wealth behind it (chapter 1). The theories from Gottfried Wilhelm Leibniz, Claude Shannon, and Alan Turing state that a machine can house information, information can be reduced to a pattern of binary sequences, and that digital computers are machines that can simulate any other machines (Universal Turing Machine). Adding in that humans are simply machines, you get AGI: A machine with a “human” mind based on mathematics. To implement this vision, computer scientists have had to answer questions about sentience and consciousness believing that AGI can know truth (Checketts uses “AGI” and “AI” interchangeably): “AI also assumes that there is a universally valid method for understanding the universe” (page 20). They ultimately settled on intelligence since it is quantifiable (e.g., IQ test). However, people quickly realized the difficulty in making AGI, and so they focused on Narrow AI: machines that can do pieces of a human mind like answering questions, interpreting medical images, etc. “The ‘objective’ claims of AI not only tend to be the purview of a privileged class within society, especially white men, but also are themselves controversial perspectives from within the Western (largely male) philosophical tradition” (page 25). The wealthy control AI.

AI promises a better world including increased wealth, but in the process, it has inscribed many societal biases (chapter 2). “AI researchers hope AI will overcome human moral weaknesses and intellectual blindspots” (page 30), but it has not; bad data in means bad data out. For example, software designed to recognize human faces categorized African American faces as those of animals (chimpanzees) and not humans, replicating horrifying racism. Furthermore, the whole endeavor of computer science is deeply gender biased – a ‘computer’ initially described a job held primarily by women; only when ‘computing’ became a man’s job was it associated with intelligence. Women are seen as less than. (See Chapter 2 for more examples.) Weapons of Math Destruction by Cathy O’Neil dives deep into the biases inscribed by machines. These examples beg the question of the moral status of machines and who is to blame for these biases? One bias not referred to much in the literature is the poor and Checketts sets out to understand their perspective in relation to AI.

To do so, one must understand the goals of AI. One promise of AI is to save money by maximizing profits through increased efficiency (chapter 3). Who sees the saved money? Not the average worker because “profit maximization depends entirely on selling products for more than the cost of producing them” (page 76). The average worker’s salary is a part of the cost of producing an item and so the goal is to minimize the cost of production, including the cost of compensating workers. Machines replace humans, leaving workers without jobs and more impoverished. The goal is to reduce humans’ intelligence to numbers to create an AI to do the task. “Artificial general intelligence, then, stands as a fetish of fetishes. It is the dream, the fantasy really, of the capitalist class. It converts all real information to numbers” (page 80).

Not only are the workers reduced to numbers, but the poor especially are seen as ‘disvalue’. The poor are a mirror to the wealthy of what they do not want to become. In a capitalist society money is valued and so the poor are disvalued. “The fate of global manufacturers replaced by automation, the charitable giving directed toward ‘AI safety’ instead of poverty alleviation, the algorithmic inscription of structural bias, environmental racism perpetuated in the struggle for rare-earth minerals and massive energy consumption, and all the countless poor who are treated as expendable in the data training process are all pressing issues brushed aside in favor of triumphal praises of the capital gains promised by AI” (page 81). Working conditions are unbearable with many large corporations like Amazon under fire for anti-labor policies. Lots of money is going into AI, but has it truly solved humanity’s problems, especially poverty? Do the poor have voices to share their concerns?

The wealthy fight back against the workers by choosing the story to tell about the poor in relation to “objective” AI: “The ‘objective’ nature of AIs that carry out these tasks on behalf of weak humans show the poor are themselves to blame… So powerful tech companies that displace, police, and impoverish the poor are considered ‘innovative’ in their use of technology (the sociotechnical imaginary), and the poor are considered morally deficient” (page 83). The story that the wealthy earned their place and the poor deserved theirs is perpetuated. This narrative maintains the divide between wealthy and poor; only if the poor are willing to play by the rules of society (set by the wealthy) are they allowed value. Policies even play into this game: Checketts argues that “every policy designed to help the poor is wrapped in requirements and policies designed to enforce bourgeois values against the poor. If the poor are more susceptible to substance abuse, US politicians will enact drug testing for welfare recipients (despite this being more wasteful)” (page 85). The poor do not get to tell their story, but have it written for them by capitalism.

Checketts aims to share the realities of poverty (chapter 4). (He shares his own story of poverty in the preface.) The “very basic definition of poverty employed in this work is scarcity or insecurity of necessary resources” (page 92). Even with this definition, Checketts struggles to tell the story because of a “Paradox: by whatever measure we might use, the worst off are bound to be those whose realities cannot be put to words” (page 94). The poor are voiceless and nameless. We only hear stories from the wealthy perspective. He continues on to state that four features help us understand poverty: lack of resources, cultivated tastes, survival mentality, and economic consciousness (page 97). The poor experience a world that is not dictated by them, but by the wealthy. Resources, tastes, mentality, and money are heavily influenced by society determining how to ‘help’ the poor, who then become even more defined by their lack of choice and short-term mentality. Surviving today right now is always top of mind, and one takes what they can afford no matter what it is: “Many poor individuals are willing to do whatever it takes to survive. Thus, property crime does track positively with economic inequality. However, violent crime does not” (page 120). These generalities do not do justice to the experiences of the poor and so other people have told their narratives through cinema including Squid Games, LadyBird, Trailer Park Boys, Parasite, etc. We need to continue hearing these stories.

From these stories, Checketts begs the question: “As the technological equivalent of the Holy Grail, does AI offer a way out of this system, or is it merely the latest tool to reinforce bourgeois epistemologies?” (page 126). He thinks that “perhaps the vision of AI fulfilling every consumerist fantasy can be reimagined with AI empowering the poor to advocate for themselves, secure better outcomes, and live free of bourgeois domination” (page 126). AI can be a powerful tool to help the broken world if used for humanity and not focused on maximizing profits.

At the same time, the poor have tried to rise up in the past (Luddism), and not only did it fail, but also it was outlawed (chapter 5). Checketts states, “AI is for the interests of the rich” (page 131). This begs the question of whether AI will result in a just outcome. Whatever the answer, it is from the perspective of those best served by capitalism and AI. However, we can look to religion to help guide us to focus on the poor. Many religions and philosophers promote helping those without means. Judaism has many mitzvot (commandments) specific to the poor including for farmers to leave any leftover or forgotten produce and the corners of their fields for the poor to glean (source: https://www.myjewishlearning.com/article/peah-the-corners-of-our-fields/). Christianity also states “Blessed are the poor, for theirs is the kingdom of God” (Luke 6:20) (page 133). Philosopher Friedrich Nietzche also states that the Bible is an inversion of the morality of the strong: “The poor is the one who is to be revered against the powerful and mighty” (page 133). Perhaps the most palpable is Jon Sobrino echoing Ignacio Ellacuria comparing the “civilization of poverty” vs. the US laissez-faire style “civilization of wealth” (page 138).

Poverty “does not seek infinite growth or acquisition of capital; it prioritizes meeting all human rights and promoting ecological and communal flourishing… The consumer is replaced by the citizen” (page 138). AI is a part of the civilization of the wealthy: “The shiny future of automation is nothing but a consumer’s fantasy, where every pleasure is somehow satisfied with no cost, where machines automatically maintain everything, where all problems are resolvable through science and technology, where our power is unbounded. It is, in a word, utopian. But it is a utopia told from the impoverished imagination of the industrial upper class” (page 139). On the other hand, the Civilization of poverty is “a civilization where the voices of the poor are uplifted, where human needs are met, and where profit is subordinate to justice is a necessary corrective to the inhumane way the poor experience the world. This view is just as utopian as the AI fantasy, but the utopia of the civilization of the poor is more authentic to the Christian tradition. This vision takes its cue from the promises God reveals to us: the last shall be first, every tear shall be dried, the nations of the world will gather as one” (page 139- 140). Religion focuses on unity and seeing humanity. Checketts argues that humanity needs to move towards the civilization of poverty to help fix the broken world.

The poor live their lives and recognize there is more to life than math and information. They do not seek to create themselves in their own image as computer scientists have created AI in their image (see table below). Whether you ascribe the mystery of life to God or anything else, AGI aims to mathematize the beautiful mystery of humanity, consciousness, and intelligence. Math cannot solve everything. The Unabomber was a math prodigy: “The mathematical mind, honed to see problems as solvable only in pure numbers, could not remain standing against the brute reality of human existence – the cruelty of child bullies, the perversity of science without ethics, … the cold mechanical calculation that hangs price tags on inestimable nature” (page 163). Calling machines human and/or granting machines rights while others do not get the same moral status (including the poor, disabled people, etc.) is problematic (conclusion). “The more that AI’s epistemological model is reinforced as normative, the epistemology of the poor is further denigrated. To say that AI is ‘conscious’, that it deserves human rights, or that it should be granted legal autonomy is to make a claim for it that is not always given to the poor” (page 177). All humans have the right to dignity, and it is not defined by intelligence.

Conclusion:

The world is polarized in many aspects and yet people who do not have choices on how to live (such as the poor) believe in things bigger than themselves. We, as humanity, need to stop trying to take over the world, and instead live and help each other. When the world fell apart during the COVID-19 pandemic, in the beginning at least, most people stepped up to help one another and did not care about any differences! Let’s go back to that, but ideally without a crisis.

Wherever you are in your journey and feelings about a topic, stop fighting! It is a good thing for us to disagree. Disagreements are how great things are made (see article: https://www.rd.com/list/petty-conflicts-changed-history/). We do not need to kill each other. Listen to each other’s stories and care for your fellow humans! We are all human beings! Poor Technology reminds us that AI is the ideal mirror for the wealthy made by the wealthy to keep their status, but the poor are a real mirror for society of the struggles we still face and need to work together to resolve with all voices heard. This book is a call to action to remember the beauty of the mystery of life and to not get hung up on how to recreate it. For a good history of AI, the poor, and their interaction, I highly recommend Checkett’s book, Poor Technology.