Debate: AI Model Training, Low-Wage Labor, And Reconstituting Societal Norms
The blame for this is on all of us, not just the AI companies
Thanks to those who have purchased a voluntary paid subscription. It has allowed me to devote more time to this blog.
Introduction
AI models are often trained using a technique called reinforcement learning from human feedback (RLHF).
RLHF allows AI systems to learn complex tasks by incorporating direct human guidance, enabling them to better align with human preferences and values.
The process of gathering human feedback to train these AI models has raised ethical concerns. Many (not all) tech companies rely on low-paid workers in developing countries to perform tasks such as labeling images, annotating objects in videos, and even composing text to train AI systems. These workers, often recruited through crowdsourcing platforms, are paid meager wages for their labor, and work in difficult conditions.
An article on The Exploited Labor Behind Artificial Intelligence highlights the plight of these workers, stating, "Unlike the 'AI researchers' paid six-figure salaries in Silicon Valley corporations, these exploited workers are often recruited out of impoverished populations and paid as little as $1.46/hour after tax.” As described in this article on AI's Hidden Workers, "between 150 million and 430 million people do such work, according to a recent World Bank estimate: they annotate images, text, and audio; create bounding boxes around objects in images; and, more recently, write haikus, essays, and fictional stories to train the sophisticated tools that could eventually replace people like me.”
Many people have expressed anger and frustration over the exploitation of these workers in the development of AI systems. Critics argue that the AI industry is built on the backs of underpaid and undervalued laborers, while tech companies reap the benefits. As noted in this article on The Exploited Labor, "The 'AI' industry runs on the backs of these low-wage workers, who are kept in precarious positions, making it hard, in the absence of unionization, to push back on unethical practices or demand better working conditions for fear of losing jobs they can't afford to lose."
I’ve heard of students and professors who refuse to use generative AI because of this.
While it is easy to express outrage at companies paying workers low wages and having them work under such conditions, the issue of low-wage labor in the context of AI model training must be understood within the broader framework of a cultural and socioeconomic system that values human labor based entirely on market principles.
In the modern economy, labor is treated as a commodity, with its value determined by supply and demand. This market-driven approach often leads to significant wage disparities, particularly between developed and developing countries but also within countries. In regions where labor supply is abundant, skills required to do the job are low, and regulatory protections are weak, wages tend to be low, creating an environment ripe for exploitation by industries seeking to minimize costs. This is true in every country in the world.
Beyond the AI Industry
The use of low-wage labor is not confined to the AI industry; it permeates many sectors of the global economy.
As consumers, we often purchase products made under conditions of low-wage labor, including clothing, furniture, and electronics. For instance, many popular clothing brands source their garments from factories in countries like Bangladesh (H&M, Gap, Lidl) and Vietnam, where workers are paid a fraction of what their counterparts in developed countries would earn.
Similarly, much of the affordable furniture available in major retail stores is manufactured in countries with lower labor costs, such as China and India.
As one Reddit user pointed out, our demand for cheap consumer goods and fast food necessitates a low-wage workforce.
Electronics, a staple of modern life, are also often produced under similar conditions. Major tech companies assemble their devices in countries like China, where labor is cheaper. The production of smartphones, laptops, and other gadgets involves intricate and labor-intensive processes, often carried out by workers who receive minimal compensation.
Our dependence on electronics also ties into another critical issue: the mining of rare earth minerals (REMs). These minerals are essential components in the manufacture of various high-tech devices. The extraction of REMs often occurs under harsh conditions in developing countries, where labor laws are lax and worker protections are minimal.
Cobalt mining, a key component in the production of lithium-ion batteries, is particularly notorious for its inhumane labor conditions. In the Democratic Republic of Congo (DRC), which produces over 60% of the world's cobalt, miners, including children, work in perilous conditions for meager wages. They face constant exposure to toxic materials, risk of injury from tunnel collapses, and long hours with little regard for safety. The lack of proper equipment and protective measures makes cobalt mining one of the most dangerous jobs in the world, with severe health repercussions for the workers involved.’ [Kids Mine Cobalt]
Cobalt is a critical element used extensively across various industries, thanks to its unique properties. One of the primary applications of cobalt is in the production of rechargeable batteries, which power a vast array of devices, including smartphones, laptops, tablets, and electric vehicles. These batteries, especially lithium-ion batteries, benefit from cobalt’s ability to increase energy density and enhance battery life and stability.
The journey of our electronics does not end at production; their disposal poses another significant challenge. Many electronic devices are discarded in developing countries, where they contribute to growing e-waste problems. Improper disposal and recycling of these devices release harmful substances into the environment, including lead, mercury, and cadmium. Workers in informal recycling sectors, often without any protective gear, dismantle electronic waste, exposing themselves to toxic chemicals. This exposure can lead to serious health issues, including cancer, respiratory problems, and neurological damage.
Prioritizing Artificial Intelligence Over Human Intelligence
As noted in our book, Humanity Amplified,”
Our Demand for Cheap Products, High Stock Market Returns, and Labor Exploitation
In today's digital age, consumers increasingly expect access to sophisticated AI-powered products and services at little to no cost.
This demand for affordable or free AI products places significant pressure on companies to find cost-effective ways to develop and maintain these services. A recent Adobe study found that 77% of consumers want brand experiences to be personalized to their needs, further driving the need for AI-powered solutions.
By employing workers in developing countries to perform these labor-intensive tasks, companies can reduce operational costs and offer their AI products at prices that appeal to cost-conscious consumers.
In addition to consumer demand for low-cost AI products, investors in AI companies, including everyday investors and those hoping their retirement portfolios perform well, also exert significant influence on corporate strategies.
To achieve these goals, AI companies are incentivized to maximize profitability and minimize costs, which often leads to the adoption of cost-cutting measures, including the use of low-wage labor.
The interplay between consumer demand for affordable AI products and investor expectations for high returns creates a feedback loop that perpetuates the focus on low-wage labor. Consumers' desire for low-cost or free AI services pressures companies to minimize production costs, leading to the outsourcing of labor to developing countries. At the same time, investors' demands for robust financial performance reinforce the need for cost-cutting measures, further encouraging the use of low-wage labor.
It’s this same societal demand for low-cost products and stock returns that are pushing companies to automate work.
Beyond Outrage at AI Companies
The exploitation of low-wage workers in the development of AI is not an isolated issue; it is a symptom of a much larger societal problem that demands our immediate attention.
In today's market-driven economy, the value of an individual is frequently reduced to their economic output and productivity. People are seen as "human resources"—just a another input in the production process that can be hired, fired, outsourced, or automated away based on market forces and the bottom line. They are commodities, disposable, and easily replaced, and they are now replaced not only by lower-wage workers but also by AIs.
Unsurprisingly, this commodification of labor dehumanizes workers and erodes the social contract. Under this paradigm, a person's worth and right to a decent standard of living are contingent upon their employment and ability to generate profits for others. Their basic needs are treated as a personal failing, regardless of the availability of jobs, rather than a collective responsibility.
As AI continues to advance at an unprecedented pace, automating more and more jobs, we find ourselves at a critical juncture. If we do not take action now, we risk creating a future in which a significant portion of the population is left without the means to support themselves and their families. The consequences of such a scenario are dire: widespread poverty, social unrest, and a fundamental breakdown of the social contract that holds our communities together.
This commodified view of human beings as merely cogs in the economic machine is antithetical to the core idea that everyone has an inherent right to a basic standard of living, regardless of their employment status or economic value to others.
To change this, we must do more than rage at AI companies. It requires a fundamental change in how we see and value one another—not as commodities, but as inherently worthy human beings with an inviolable right to a dignified existence. While daunting, this shift in consciousness is essential if we are to build a truly just and caring society in an age of ever-advancing automation. The alternative is to continue drifting toward a dystopian future of extreme inequality, precarity, and desperation for the many who are left behind by an increasingly unequal and unforgiving economic order. It will not be a utopian future that individuals like Marc Andreesen and Elon Musk (‘universal high income”) identify.
The idea that we will look out for each other and sustain a world of high income and quality of life for all without some fundamental change in social norms is absurd.
If we are going to achieve anything like the world AI utopians envision, it will take more than technological advances. It will require us to invest at least as much in human intelligence as we do in machine intelligence, and in the status quo, we are increasing investment in machine intellligence and decreasing investment in human intelligence
It will require us to reconstitute a new ethical relationship with each other, one that prioritizes empathy, compassion, and solidarity over profit and efficiency. We must fundamentally shift how we think about work, wages, and the role of technology in our lives.
This requires a collective effort from all sectors of society—governments, businesses, and individuals alike; you cannot outsource your responsibility. We must strive to create a society in which the benefits of AI and automation are distributed equitably, ensuring that no one is left behind.
Questions
Such a reconstitution requires action by all individuals. If you are upset with AI companies, what do you think of the following?
If you won’t use AI because of worker exploitation, have you also stopped using other products that arguably exploit workers?
Do you regularly pay more for products that are produced through “fair trade,” to help workers?
Will you abandon your electronics in to stop practices that exploit workers and the environment? Reduce your upgrades?
Are you willing to pay more for your AI subscription so laborers can be better paid? Are you willing to pay more for food? Clothes?
Do you pressure companies to pay their workers more, even if it means higher prices?
Are you ok with lower stock returns as long as workers are paid better?
Before purchasing clothing, do you try to determine where it was made and that workers were properly paid?
Do you tip as much as you can?
If you are doing at least reasonably well, are you willing to pay higher taxes to redistribute income to the less fortunate?
