People behind the system
Ghost workers who work in the shadows of APIs
In SciFi movies, computers are just like humans. Ironman’s computer recognizes the face of a villain and retrieves personal data within just a few seconds. Growing up with these movies, most people are used to the ideal futures of computing, magic where technology happens just by itself.
However, that is not true in the real world. Computers still need a touch of humans to work properly. For instance, AI has to go through a training process that includes human labor labeling the information. There are always people who feed and complement the system. In Gray & Suri’s book ‘Ghost work(2019)’, they called these workers ‘Ghost workers’ who are invisible and works in the shadow of APIs.
Today’s artificial intelligence can’t function without humans in the loop. (Gray & Suri, 2019, p.5)
These workers are situated between computers and humans. They cannot be like computers which are always objective, accurate and unbiased. Also, even though the engineers would expect human to bring creativity and innovation to work that CPU lack (Gray & Suri, 2019, p.10), it is not possible for them to be ‘humane’ since they are under the supervision of a ‘requesters’ and often forced to follow their values.
A good example of this is discussed in the article ‘Social Media’s Silent Filter’ written by Sarah Roberts. In one of the examples she mentioned, the CCM worker who screens contents uploaded on social media had to let the videos on the Syrian war stand in the platform but had to flag videos regarding the drug wars of Nothern Mexico. Even if the workers felt they want to do the opposite way, they had to follow the codes of conduct of the requesters as they might not want to harm their ratings which would affect their access to future tasks.
As in this example, the requesters consider the workers as a part that complements the system but not a distinct individual. It is easy to notice this by looking at the MTurk system where workers are identified with numbers and mostly left anonymous. There seems no room for any personal opinion or creativeness of workers could play. In other words, the on-demand platforms and the structure of the ghost works are treating people as a part of a computer system, making them lose their agency.
Also, the legal status for them has not been established and the fact puts the workers a more vulnerable position where they cannot be protected from mental and physical damages from the contingent works. As discussed in the article by Sarah, there are many cases the workers were exposed to heinous examples of abuse and violence which resulted in permanent damages for some. (Sarah, 2019)
Labor is not just about earning money. Labor should give people a feeling of connection, motivation and also self-improvement. Some people can also find the reason for their living at work. However, the way the on-demand platform is designed is making people more focused on the financial part of work, making them forget about what they can get from working and be just brain-dead repeating a simple task wanting more money.
The generic intention of having a human computation was to rely on the person’s creative capacity and agency. (Gray & Suri, 2019, p.5). However, the agency doesn’t seem like working anymore in the current system of the ghost work platforms. Like what is discussed in Gray & Suri’s book, computers will constantly need human labor as the boundaries of automation would keep pushing and there will be constant ‘last mile’ that needs to be filled with humans. (Gray & Suri, 2019, p.16). Therefore, it is crucial to address how to consider these workers, not as a module of computer, but a human who has agencies and dignities.
