READY FOR GRADING

Ghost work was a term coined by anthropologist Mary L. Gray and computer scientist Siddharth Suri in their 2019 book, Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass. The authors' intended use of the term 'ghost work' is to call out work conditions where the value of the person providing momentary service is erased.

Definition[edit]

edit

Ghost work does not describe the work itself, but the conditions of work. For companies that use content-control software to filter inappropriate material, ghost work is necessary when Artificial Intelligence (AI) cannot recognize insights that humans can, such as discerning between images of traffic lights and crosswalks[1]. It focuses on task-based and content-driven work that can be funneled through the internet and Application programming interfaces (API's). This kind of work can include labeling, editing, moderating, and sorting information or content.

 
An example of CAPTCHA which tests whether users are humans or robots

Ghost work is also considered work that can be done remotely (so long as there is internet access) and on a contract-basis. It's an invisible workforce, scaled for those who desire full-time work and for those who pick it up whenever they have the time. Though it can work position-independent through the internet, there are data factories in China that mine “the Saudi Arabia of data” by parsing and cataloguing to make data useful and then assemble the foundation of the nation’s AI ambition. The core characteristics of ghost work are considered low-wage, disposable, and menial. Ghost workers are the low-tech part of the high-tech production, as the construction workers in the digital world.

A benefit of ghost work is flexible hours due to the worker choosing when they complete a task, making it an appealing option for those in between jobs or those in need of side work. However, with the promise of flexible hours and endless tasks, companies can potentially undervalue, under appreciate or under compensate workers. The workforce today is beginning to adapt more towards this labor style, similar to Uber and Lyft drivers, as opposed to the standard 9-5 workday.

In contrast to peer production that emphasizes the community spirit and co-work on open source products, ghost work tends to be benefit-driven.

Ghost work differentiates from gig work or temporary work specifically as it is task-based. While gig work includes more of a general platform work, ghost work emphasizes on the software or algorithm aspect of assisting machines to automate further. Through labelling of content, ghost workers teach the machine to learn as defined by Gray and Suri “human labor powering many mobile phone apps, websites, and artificial intelligence systems".

Examples of ghost work

edit

Amazon (company) is the most notable example for this type of work; for, as the retailer grew, the company realized that they would need to constantly post products, verify product photos, create product captions, and more. In addition to these tasks, Amazon also needed an army of people to update book reviews dating back to 2005. Consequently, the website Amazon Mechanical Turk (otherwise known as MTurk) was created for "crowdworkers" to pick up and complete posted tasks. Once these tasks were completed, the person who completed it would be paid.[2] Amazon also charged a small surcharge to match posters with those who had certain qualifications to complete the projects and tasks. This allowed almost anyone to go on and find work. This platform allows for easy and inexpensive participation among workers, particularly young individuals.[3]

The devaluation of momentary service[edit]

edit

The concept of hiring on-demand workers is not radically new. By the late 1800s, Lowell mills paid farm families to hand-fashion cloth pieces into shirt flourishes that were still too delicate to churn out on the factory floor. Equivalently, today's companies hire on-demand workers to test the latest ranking, relevance, and crawling algorithms of their search engines so they can be perfected.

Contingent work was further devalued by culturally loaded notions surrounding what constitutes learned profession or "skillful" work, as well as which workers deserved or needed full-time jobs. A contract worker helping with a speculative education software package, that may or may not ship, could be written off, in part, due to gender, skin color, nationality, professional training, physical location, or all of the above.

Those doing on-demand jobs today are the latest iteration of the expendable ghost work. On the one hand, they are necessary in the moment, yet undervalued as mundane and unsuccessful.

External image[edit]

edit

The computer science world, including a variety of tech companies, are invested in producing the image of technological magic. Amazon Mechanical Turk hides the people involved in the production of so called magic, whose visibility would otherwise obstruct favorable perception. The following isn't only aimed at public image of the company but also at investors, who are significantly more likely to back businesses built on scalable technology, not unwieldy workforces demanding office space and minimum wages. In addition, there exists a pervasive belief among engineers that these workers are a stop-gap until AI can replace or do without them, which inevitably leads to their vital contribution being devalued. Despite the belief, the market for ghost work doesn't show apparent signs of declining. If anything, it's contrary.

Bibliography

edit
  1. ^ Gray, Mary L. (2019). Ghost work : how to stop Silicon Valley from building a new global underclass. Siddharth Suri. Boston. ISBN 978-1-328-56624-9. OCLC 1052904468.{{cite book}}: CS1 maint: location missing publisher (link)
  2. ^ Difallah, Djellel Eddine; Catasta, Michele; Demartini, Gianluca; Ipeirotis, Panagiotis G.; Cudré-Mauroux, Philippe (2015-05-18). "The Dynamics of Micro-Task Crowdsourcing: The Case of Amazon MTurk". Proceedings of the 24th International Conference on World Wide Web. WWW '15. Florence, Italy: International World Wide Web Conferences Steering Committee: 238–247. doi:10.1145/2736277.2741685. ISBN 978-1-4503-3469-3.
  3. ^ Huff, Connor; Tingley, Dustin (2015-07-01). ""Who are these people?" Evaluating the demographic characteristics and political preferences of MTurk survey respondents". Research & Politics. 2 (3): 2053168015604648. doi:10.1177/2053168015604648. ISSN 2053-1680.