By Dr Sarah Cant
Seismic change often arrives on cat’s paws, argued sociologist, Ulrick Beck, in his book Risk Society; its pervasiveness and persuasiveness can take us by surprise.
This metaphor certainly captures the advent of the age of artificial intelligence (AI). Whilst AI has a relatively long history – Alan Turing, famously proposing his test to measure machine intelligence in the 1950s, the breakthroughs in deep learning and neural networks that allowed computers to recognise images through to large language models that centred generative AI into our everyday lives, occurred only in the last decade. ChatGPT arrived without much fanfare in 2022, and quickly but quietly transformed our human made world in myriad ways.
It is undeniable that AI has altered the ways that we live, work, learn, teach, research, think, interact, communicate, consume, love and wage war. AI brings opportunity and hope, operational efficiency, and breakthroughs in medicine, for instance, but also indicates risk and uncertainty. Some AI bosses warn of the end of humanity, some recruiters fear that it is making applicants ‘stupider’, and there are widespread concerns about the environmental impact of powering its global infrastructure. Certainly, AI is now our go-to resource, whether for information, email construction, or even friendship. Therefore, sociology is imperative to navigate both the intended and unintended consequences of this digital revolution and to give us the tools to shape its future. As Joyce and Cruz (2024) argue, ‘the sociology of AI moves us beyond headlines and hype to spark innovative sociological work toward equity and justice’.
Bestowing a critical disposition
In How to be a Social Researcher, we argue that sociologists uniquely have the skills to interpret and assess data. This is now essential expertise when reading the results of AI generated searches. AI depends on existing archives which it uses to pull together reams of data in a fraction of a second, which it can then summarise effectively. This is a marvellous capacity and one that I use regularly. However, there remains a deep sociological issue. All knowledge is socially constructed. By this, I mean that all our ideas are shaped by social circumstances – our thinking always reflects politics, economics, cultural mores, and societal divisions. Therefore, it follows, that the archives that AI draws upon cannot be regarded as simply objective or valid or reliable. Rather, the archive contains societal prejudices and stereotypes, and this raises deep ethical questions. Let me give you an example. I asked AI to generate images of the working class – and, in less than a second, I had an array of pictures, but they were dominated by depictions of single mothers, smoking and worklessness. Sociology gives us the critical disposition to challenge this simplistic and stereotypical depiction. While AI is now integral to our social lives, sociologists are needed more than ever to reveal any inherent biases and to question what might be easily be taken for granted to be truths.
Asking the right questions
Indeed, in How to be a Sociologist, we show that the sociological imagination, the ability to see connections between personal concerns and wider social structures, allows us to ask deep questions about the social world and challenge taken for granted assumptions. These questions can be applied to AI. Sociologists ask probing questions such as: who or what is in control?; who is accountable? who benefits?; who is excluded?; what counts as knowledge?; how do people use AI in everyday life and work? To provide answers to these questions, we can draw on our rich sociological theory.
Let’s look at some recent work. Digital sociology has long enabled us to explore how new technologies shape our identities and institutions and, drawing on the work of Foucault, think about the pervasive potential of AI to surveil our private and public lives. In turn, it details how structural inequalities can be reproduced by AI. For instance. Raji et al (2020), showed that facial recognition technology misclassifies marginalised groups, and indicates the endurance of the W.E.B Dubois’ colour line in reproducing inequality. Feminist researchers have shown that AI also embeds gender bias. Sociologists are active in identifying these inequalities and Silicon Valley are training AI practitioners to devise fairness metrics for AI systems. As another example, Marxist theory and the sociology of work helps us think about the rise of robots and machine learning and asks questions about alienation and immiseration. At the same time AI is making working easier in many ways for workers, and good sociology needs a nuanced and balanced discussion.
Sociological Intelligence
There are widespread concerns about the impact of AI on schooling that include worries about cheating, fears that teachers might be replaced by chatbots, and widespread anxiety that we are losing our capacity for human creativity and expression – what Marx referred to as our ‘species being’. Yet sociology students are uniquely equipped with the skills and disposition to engage with AI in a careful, considered and reflexive way. AI is here to stay and offers many opportunities. The sociological imagination provides the capacity to use AI critically and creatively, moving beyond passive consumption towards meaningful interrogation. Moreover, a sociology of AI is emerging and, as the next generation of sociologists, students will be central to shaping its development, direction and impact.
Dr Sarah Cant is Professor Emerita at Canterbury Christ Church University. She is passionate about the value of studying sociology in schools and at university. With her colleague, Jennifer Hardes Dvorak, she has written two books aimed to inspire the next generation of A Level sociologists. These books showcase the contemporary insights that sociology affords, the value of sociological research methods, and the transferable skills that sociology bestows.