AI is not so artificial after all
The beginning of the Industrial Revolution in the 18th Century was a significant era for the automaton. This was also the era in which the Mechanical Turk made its appearance: an automaton chess player that could defeat human chess opponents and was said to have been a conscious machine. This seemingly marked the birth of the phenomenon we now know as artificial intelligence.
The Mechanical Turk made chess players think they were competing against a machine, and it scared people as much as it entertained them. No one, not Napoleon Bonaparte nor Benjamin Franklin, could checkmate it. In the end, it turned out that the chess skills of the Mechanical Turk were the results of a human cleverly hidden in a secret compartment.
Now, fast forward to todays’ economic revolution, which is disrupting the economy, companies, labor markets, and our lives in a way not seen since the Industrial Revolution. In the midst of this revolution, we are still being fooled by a modern version of the mechanical turk.
Picture this scenario. As a consumer, you use the Uber app on your smartphone and an Uber driver responds. But neither you nor the driver knows that your meeting hinges on the work of another human. The hidden secret compartment.
The moment a computational process (machine) fails — for example if the facial recognition technology of Uber’s Real-Time ID Check fails to recognize your driver because he shaved his beard — a human far from Silicon Valley steps in to fill this gap. After uploading the job route from Uber to a platform aptly titled The Mechanical Turk, this person is given a short amount of time to determine whether your driver is the same person as his registered ID. This means another human is an invisible yet integral part of your ride.
This is an example of a task, a paid task, known as ghost work.
AI’s little secret
In the age of artificial intelligence (AI), work done by humans is not disappearing. It is being hidden.
It is undeniably true that robots are on the rise but most automated jobs still require human expertise. This is not new from a historical perspective and began as early as the Industrial Revolution with a phenomenon known as piecework. A prime example of this was textile mills that could produce a shirt and even mass produce t-shirts but couldn’t add a bow or a flourish of some kind to the shirts. This industry relied on family farms around the industrial cities to complete that final piecework. Ghost work operates on that same principle.
And it is everywhere. The truth is that the services provided by companies like Amazon, Google, and Uber only function properly because of humans working invisibly in the background. So in essence, ghost work can be described as the final piecework AI could not do: AI’s little secret.
Lifting the curtain
From being content moderators on Twitter to helping Alexa understand regional dialects, the tasks of ghost workers can comprise a diverse set of areas, as long as they minimize the cost and time of machine learning development.
The blend of human creativity and code is what crowdsourcing marketplaces such as Amazon Mechanical Turk (yes, it’s quite remarkable that the name for this crowdsourcing marketplace was inspired by the hidden secret compartment of the Mechanical Turk) and Figure Eight propose. These platforms allow companies such as Uber, Google, and YouTube to remotely hire workers worldwide to do the tasks that AI cannot do. In many cases, it allows companies to sell their services as if they are being delivered by the magic of AI.
Ghost workers come from all walks of life, from young mothers to minorities to recent grads. They choose to do ghost work because of the flexibility it offers and the liberty to work from home. The problem is that there is no economic model for this type of employment. The moment Google, Uber, or Amazon needs humans to fill the gap, they need it fast. Traditional hiring methods will not do because we, as end-users, will not tolerate any delay waiting for our Uber driver to arrive.
In addition, there is little transparency about the reason the work of ghost workers gets rejected or what their work is contributing to or building. So the invisibility of these workers results in their marginalization. The cruel irony is that these are the people who make sure we can safely navigate the internet and have on-demand services.
The latter is particularly essential for the public interest, primarily for our digital future. Because even though these workers are in the background, they have an immense responsibility as the arbiters of online content. They are partly in charge of training algorithms to make crucial decisions about good or harmful content. If they are biased, the algorithm they are training will also become biased, which will impact our digital future.
Out of the shadows of AI
It is clear that there is nothing obvious about what we humans do and say, i.e., our cognitive abilities, and the world should know that human labor powers present-day AI.
The overall goal of AI is to build a computer system that mimics human-like abilities such as learning and problem-solving. However, unlike humans, AI has not been shaped by natural selection — only a human can make a snap judgment in the heat of the moment with no prior information. That is a world of work that we cannot see.
The on-demand service industry is turning into task-based work. Humans intervene where AI fails and where they excel: react quickly and spontaneously and conduct complex communication. These are tasks that are technically hard for AI to solve. So, in ignoring the value of the humans behind AI, we devalue the work they do when they step in and deliver that moment of intervention to help us out.
It is clear that these workers contribute significantly to the AI revolution. Thus it is unethical and unfair to discredit or devalue their contributions, value, and presence in technology. Also, because of their role in shaping various aspects of our lives, it is our duty to bring them out of the shadows of AI.
The overall goal of AI made me think: perhaps, as we train machines to become more human, are we making humans work more like machines?
Gabriella Obispa is a guest writer for Profound. She is a master’s student majoring in International Technology and Law at the Vrije Universiteit Amsterdam. As a feminist and ‘woman in tech’, she is committed to the empowerment of women and diversity within the tech scene.