Methodology

1.3.1 Original contribution

As it has emerged from the literature review, many scholars from different fields have discussed the benefit of providing a historical contextualization to the analysis of the impact of AI and big data, as well as the need to recognise the central role of culture in the development and adoption of new technologies. While Module C takes advantage and builds upon this vast and relevant scholarship, it will also provide an original contribution to this debate. This is due to the two peculiar approaches distinguishing the KT4D project from similar existing academic endeavours and that are:

• First, the project’s focus on knowledge technologies and the importance, discussed in the previous sections, of this open and novel definition that identifies a category distinct from the more commonly used ones of media and information technologies. This is because the project recognises the process of knowledge creation and sharing as central to democratic participation.
• Second, this Module adopts a systematic approach to the comparative analysis of past and present knowledge technologies from a cultural studies perspective. As described in the previous sections, there is a growing number of scholars historically contextualising AI tools and systems and drawing parallels with past technologies. However, their focus is usually limited to one issue or technological application at a time, like in the already mentioned comparison between deepfakes and photographic manipulation of the 1920s. What Module C aims to accomplish, instead, is a comprehensive map of past and present knowledge technologies so as to identify general trends, divergences, and similarities. The overarching theme threading these case studies together is the definition of knowledge technologies to which they all refer to.

3.3.2 Establishing precedents: not a list of technologies, but a list of issues

To adopt a definition of technologies as systems (see section 1.2.3.1), and of knowledge technologies as systems specifically involved in the process of sense-making, it means that past and present examples of KTs can only be understood historically and contextually. For this reason, the analysis developed in Module C will not consider a list of specific examples of past KTs (e.g. the printing press, television, Web 1.0) and then compare them one by one to advanced KTs (AI and big data). This is because an approach of this sort would assume that the technological element is preponderant compared to the human one (something that is closer to the definition of information rather than knowledge technologies, see section 1.2.1.2), a hypothesis that many scholars has refuted, as shown in the literature review, and to which Module C subscribes.

Furthermore, this approach would lead to erroneously consider KTs as tools rather than systems and would impose an abstract and essentialist idea of what each technology is, which is something that Gitelman, among others, criticises and warns against:

So it is as much of a mistake to write broadly of ‘the telephone’, ‘the camera’, or ‘the computer’ as it is ‘the media’, and of – now, somehow, ‘the Internet’ and ‘the Web’ – naturalizing or essentializing technologies as they were unchanging, ‘immutable objects with given, self-defining properties’ around which changes swirl, and to or from which history proceeds. Instead, it is better to specify telephones in 1890 in the rural United States, broadcast telephones in Budapest in the 1920s […]. Specify is key. Rather than static, blunt, and unchanging technologies, every medium involves a ‘sequence of displacement and obsolescences, part of the delirious operations of modernization’, as Jonathan Crary puts it. (8)

Following Gitelman’s recommendation, Module C will develop its comparative analysis of past and present KTs by assuming that something such as ‘television’ or as ‘chatbots’ does not exist, instead there are only historically and culturally realised interactions between people and versions of these technologies. Accordingly, the study will focus on similar issues across time and across technologies that are relevant to contextualise AI and big data (see Section 1.4.1). Indeed, to identify relevant and reusable patterns from past interaction with KTs (either as cautionary tales or as virtuous examples), Module C will first identify people’s needs, fears, hopes, problems around AI and big data, drawing from the insights offered in Module A and B, and then look for similar entanglements in past interactions with KTs, without incurring in misleading generalizations such as that ‘AI is the new printing press’.

The initial framework adopted in Module C to map suitable case studies considers three main categories: agency, creativity, and identity, which subsume the two aspects of trust and free will discussed in the analyses of Module A and B:

• Agency pertaining the process of knowledge access and sharing. This includes issues concerning consciousness, intentionality, free will, and autonomy. This category addresses how KTs have always been used to manipulate people through propaganda and social control, but, at the same time, they have also been used to democratise access to information and to support liberation movements;
• Creativity pertaining the process of knowledge creation. This deals with two opposing views of KTs as capable of threatening people’s capacity for creativity, understood as a quintessentially human trait, and thus limiting their freedom of expression, or, at the opposite, as tools relieving people from boring menial tasks, or even offering opportunities to further enhance their creativity;
• Identity pertaining the process of knowledge acquisition. This category focuses on the link between KTs epistemology and people’s understanding of their role and place within their community. As KTs can lead to a more truthful, effective expression of one’s identity and thoughts, those same technology my pose a threat as they can enforce stereotypes, identity-based discrimination, or simply disrupt social hierarchies and cultural norms essential to people’s identity building.

As it emerges from this framework, the analysis developed in Module C will not only consider the threats posed by AI and big data, but will also examine the opportunities that these technologies present and will do so by identifying historical precedents of how people have leveraged the power of KTs for good. In doing so, it is worth considering Mike Ananny’s interpretation of the double role and function of algorithms used in machine learning and AI systems. Ananny states that “Algorithms are both ‘traps’ that sequester people in particular cultural worldviews, and ‘societies’ that transform how ‘people interact, associate, and think.’ They simultaneously give people options for what to do, and signal what people are expected to do and what most people do” (6). Understanding not only algorithms but KTs in general as both ‘traps’ and ‘societies’ helps us recognizing how these tools and systems have the power to hamper democratic participation and personal realization, but can also be enabler of positive change and serve the needs of society as a whole. Indeed, what Ananny writes about algorithms “creating descriptions of the world that people use to reflect upon their identities, communicate with others, and create public life” (6), also suits our definition of KTs intended as cultural systems for shared meaning-making.