IMPACT COMMUNICATION
AND INNOVATION
AND INNOVATION
During this past month, themed as LGBT Pride Month, the blog Sci-Fi Interfaces dedicated a series of articles to the issue of gender in the development of characters representing artificial intelligence. Although it does not aim to correlate with the state of the art in the scientific realm, it is possible to observe how fiction and reality are in a constant exchange of values and ideals that are represented on cinema screens and also brought into the laboratory.
With a subtitle that translates to “watch less science fiction, start using it,” the blog highlights the importance of gender as a methodology and parameter for thinking about the issues of reality and the future we are building by presenting quantitative research that seeks correlations in the fictional universe to reflect on the context of our society. Among the topics covered in this series of articles are notions of subservience according to the AI's gender, categorization of intelligence types, embodiment, the distribution of benevolence, as well as notions of free will and subservience.
Moving between retro and contemporary references, the reflections proposed by Christopher Noessel are based on a methodology that attempts to quantify in order to better defend his thesis on how the gender of artificial intelligence is presented in pop culture or science fiction as a whole— in a specific post, the author details his evaluation method in depth. For the “data nerds,” as he refers to them, it is a feast for analysis.
With this, Noessel was able to map, for example, what the main physical representations of an artificial intelligence are in audiovisual science fiction. Among his conclusions is the fact that when an artificial intelligence is represented as being equal to a human being, for example like the androids in Westworld, it will inevitably be represented as a man or a woman. When the AI is designed to show no gender, however, it usually ends up having a body that is more mechanized and lacks anthropomorphic shapes, as is the case with the robot TARS in the film Interstellar.
When Noessel takes into account artificial intelligences without a body, it is in these cases, however, that the predominance is for a female voice, thus suggesting that it is a female-gendered entity. With the exception of examples like JARVIS from the Marvel universe, Noessel was able to list at least eight cases of bodiless AIs with female voices and mannerisms: Friday (Avengers: Age of Ultron), Coach (Black Mirror, episode "Hang the DJ"), Samantha (Her), VIKI (I, Robot), Gipsy Danger (Pacific Rim), Sibyl (Psycho-Pass: The Movie), Karen (Spider-Man: Homecoming), and Axiom (WALL-E).
This is a predominance that is also seen in practice, when experiments reveal that people feel more comfortable with a female voice in a robot or artificial intelligence, as well as tending to interact with digital assistants that are women. These are old, analog customs and precepts that are transported to the digital realm and to a new era which, on the other hand, needs to break the problem of sexism that remains intrinsic to technology, for example by using technology itself as a way to suppress this flaw.
Further on, Noessel also addresses the issue of gender and subservienceTo this end, the author divides characters into four categories: AIs with free will, those that improvise, those that are reluctant, and those that are slave-like. Given this, Noessel realized that a robot with free will is more likely to be portrayed as male than female or any other gender. In the case of an AI that has a master but is free to improvise, it is possible for the machine to be agender or gender-fluid. As for female artificial intelligences, there is apparently no significant disproportion between subservience and obedience—in reality, when it comes to an enslaved AI, it will likely not even have a gender or an anthropomorphic body. In this sense, Noessel's conclusion is that while the male AI represents free will, the machine represents subservience.
On the other hand, when Noessel addresses the issue of good and evil in artificial intelligences, it becomes clear that male AIs tend to be extreme (very good or very evil), while female AIs have more nuance and are closer to neutrality. As for genderless AIs, they are likely to be somewhat evil or even use their gender fluidity as part of a deception strategy—as is the case with Skynet, for example.
However, a second trend observed by Noessel is that when AIs have bodies biologically similar to women, they are more likely to be evil, but not excessively so. Furthermore, the use of female pronouns or voices has also become a way to neutralize the perception of evil or good in a machine, since the machine itself is already capable of causing unease (hence the same logic behind the use of female voices in digital assistants or social interaction robots).
Finally, when we talk about the masters and creators of these artificial intelligences, we realize yet another scenario: although we can say that more subservient AIs are mostly agender machines and AIs with free will tend to be male, this does not necessarily influence how they relate to their master or creator. What Noessel observed is that when there is a female creator, it is much more likely that her AI will be agender— and it is worth noting that there are very few women portrayed as scientists creating AIs. In the case of male creators, it is much more likely that their creations will have a defined gender and that they will be female, although Noessel notes that if we take a male artificial intelligence into account, there are simply no examples of female creators—these will only be men or they will not have a defined gender.
By mapping this type of content, Noessel helps us better understand the zeitgeist of our culture and how we are representing these archetypes in the fiction that inspires us and exchanges influences and visions of the future with us. Reflecting on this data gives us a better perception not only of narrative paths for fiction, but also of how we are translating our reality based on valuations driven by gender and its stereotypes. In this sense, the research helps us clarify not only quantitatively what pop culture is highlighting, but also a symptom of what is happening in the tech industry and what adjustments we could make based on these speculative scenarios.
