It is fair to say that there is growing concern about the use of artificial intelligence in almost all facets of our lives. In universities, a lot of this worry centres on the potential use of AI by students and colleagues to produce work without effort or meaningful understanding. I wanted to know what this would look like, so I asked ChatGPT to ‘explain the potential of AI in medical sociology’. I did this not knowing whether to expect something of a quality that could be crafted into a peer-reviewable academic paper, or a load nonsense. What I got was neither. The answer it gave was interesting, if a little arrogant on its own centrality in the future of the discipline. It was also, perhaps, indicative of public information and views on sociology more broadly.
ChatGPT offered a rather self-rewarding and grandiose start to its response:
Artificial intelligence (AI) has the potential to revolutionize the field of medical sociology by providing new insights into the social determinants of health, improving patient care, and enabling more efficient and effective healthcare delivery.
Revolution aside, I was surprised by the bots ability to identify medical sociology as a discipline, though it did increasingly focus on medicine rather than sociology. Though the bot truly grasped the concept and significance of the social determinants of health. There is some irony in the bot recognising this, while being subject to critique and concern about algorithmic bias enacted by AI already used in healthcare.
Chatting to a robot informed me of how AI is currently used in health, particularly public health and health predictions. After outlining its revolutionary potential, the bot summarised key applications of AI in medical sociology, which included predictive analytics, natural language processing, precision medicine and medical imaging – with sociology as an optional add-on.
The first potential application, predictive analytics, was explained as below:
AI can analyze large datasets of patient health records to identify patterns and risk factors for specific health outcomes, such as chronic diseases or mental health conditions. This information can be used to develop more targeted prevention and intervention strategies that take into account the social determinants of health, such as income, education, and neighborhood conditions.
This is a relatively superficial outlook drawing on public health-style monitoring. It replicates the insistent focus on individual behaviours and responsibilities relating to health.
The ordering of the paragraph also reflects the structure of how many health interventions or trials are ordered: individual risk factors and behaviours and targeted responses, with social sciences and social inequalities often included as an afterthought.
In responding to my original request, ChatGPT focused on known areas of AI usage, and did not draw on any of the traditional sociological theories that I had anticipated, nor that it suggested when I asked for it to outline key sociological theorising*. Instead, it treaded on the safe roads that AI has been used for years, such as Natural Language Processing, with the addition of acknowledging the significance of societal factors on health and where we might access that information. Though AI draws on existing knowledge, the idea below for using NLP was a surprise to me:
Natural language processing (NLP): NLP can help to identify and analyze social factors that may impact health outcomes, such as language barriers, cultural norms, and social support networks. By analyzing patient feedback and social media posts, for example, AI can help to identify social determinants of health that may not be captured in traditional health record.
Like many researchers, I have used social media to recruit participants, as a platform for discussion and seen others use it as a route to behaviours change, including health related behaviours. The chatbot goes on to outline its value in analysing medical images, following the same pattern of focusing on medical outcomes, before adding that ‘AI can also help to identify patterns in medical imaging that may be linked to social determinants of health, such as environmental toxins or occupational exposures’.
The bot loses its way after this, focusing exclusively on health outcomes with no link to sociology or any social science at all, getting even more confused in its conclusion:
Overall, AI has the potential to enhance the understanding of social determinants of health, improve patient care, and enable more efficient and effective healthcare delivery in medical sociology. However, it is important to ensure that AI is used ethically and that its potential biases are identified and addressed to ensure equitable healthcare for all.
It seemed rude not to follow up on a mention of ethics, so I asked the bot whether it was appropriate to use AI in the sociology of health and illness. The bot provided a reasonably balanced response, though illustrated the limited understanding artificial intelligence has in understanding what sociology is, or what sociologists do.
Yes, it can be appropriate to use AI in the sociology of health and illness, as it has the potential to provide valuable insights into the complex social factors that contribute to health outcomes. However, it is important to use AI in a thoughtful and responsible manner, taking into account issues related to privacy, bias, and equity.
So, do we need to worry about AI writing papers for deceitful academics in the sociological discipline? Not yet. Though this isn’t the most evolved form of AI, it is one of the most readily available and accessible. In every paragraph, despite the wealth of information available about the sociology of health and illness available, AI repeated recognisable misunderstandings and misinterpretations of sociology. It suggests that instead of fighting the growth, development and use of artificial intelligence, we should perhaps be continuing to promote and advocate for the growth, development and use of medical sociology.
*The eight sociological theories identified and outlined by the bot were:
Structural-Functionalism: Conflict Theory: Symbolic Interactionism: Feminist Theory: Postmodernism: Critical Race Theory: Social Exchange Theory: Rational Choice Theory:
Mental Health Cultures
Apr 17, 2024
Will a controversial new data platform change how the NHS “thinks”?
Apr 3, 2024
Mental health services in trouble: a pregnant psychotic patient
Mar 27, 2024
The NHS in Late Soviet Britain
Mar 20, 2024
Hospital privatisation is linked with worse quality care for patients – new research
Mar 6, 2024
Artificial Intelligence and Medical Sociology
by Jen Remnant May 17, 2023It is fair to say that there is growing concern about the use of artificial intelligence in almost all facets of our lives. In universities, a lot of this worry centres on the potential use of AI by students and colleagues to produce work without effort or meaningful understanding. I wanted to know what this would look like, so I asked ChatGPT to ‘explain the potential of AI in medical sociology’. I did this not knowing whether to expect something of a quality that could be crafted into a peer-reviewable academic paper, or a load nonsense. What I got was neither. The answer it gave was interesting, if a little arrogant on its own centrality in the future of the discipline. It was also, perhaps, indicative of public information and views on sociology more broadly.
ChatGPT offered a rather self-rewarding and grandiose start to its response:
Revolution aside, I was surprised by the bots ability to identify medical sociology as a discipline, though it did increasingly focus on medicine rather than sociology. Though the bot truly grasped the concept and significance of the social determinants of health. There is some irony in the bot recognising this, while being subject to critique and concern about algorithmic bias enacted by AI already used in healthcare.
Chatting to a robot informed me of how AI is currently used in health, particularly public health and health predictions. After outlining its revolutionary potential, the bot summarised key applications of AI in medical sociology, which included predictive analytics, natural language processing, precision medicine and medical imaging – with sociology as an optional add-on.
The first potential application, predictive analytics, was explained as below:
This is a relatively superficial outlook drawing on public health-style monitoring. It replicates the insistent focus on individual behaviours and responsibilities relating to health.
The ordering of the paragraph also reflects the structure of how many health interventions or trials are ordered: individual risk factors and behaviours and targeted responses, with social sciences and social inequalities often included as an afterthought.
In responding to my original request, ChatGPT focused on known areas of AI usage, and did not draw on any of the traditional sociological theories that I had anticipated, nor that it suggested when I asked for it to outline key sociological theorising*. Instead, it treaded on the safe roads that AI has been used for years, such as Natural Language Processing, with the addition of acknowledging the significance of societal factors on health and where we might access that information. Though AI draws on existing knowledge, the idea below for using NLP was a surprise to me:
Like many researchers, I have used social media to recruit participants, as a platform for discussion and seen others use it as a route to behaviours change, including health related behaviours. The chatbot goes on to outline its value in analysing medical images, following the same pattern of focusing on medical outcomes, before adding that ‘AI can also help to identify patterns in medical imaging that may be linked to social determinants of health, such as environmental toxins or occupational exposures’.
The bot loses its way after this, focusing exclusively on health outcomes with no link to sociology or any social science at all, getting even more confused in its conclusion:
It seemed rude not to follow up on a mention of ethics, so I asked the bot whether it was appropriate to use AI in the sociology of health and illness. The bot provided a reasonably balanced response, though illustrated the limited understanding artificial intelligence has in understanding what sociology is, or what sociologists do.
So, do we need to worry about AI writing papers for deceitful academics in the sociological discipline? Not yet. Though this isn’t the most evolved form of AI, it is one of the most readily available and accessible. In every paragraph, despite the wealth of information available about the sociology of health and illness available, AI repeated recognisable misunderstandings and misinterpretations of sociology. It suggests that instead of fighting the growth, development and use of artificial intelligence, we should perhaps be continuing to promote and advocate for the growth, development and use of medical sociology.
*The eight sociological theories identified and outlined by the bot were:
Structural-Functionalism: Conflict Theory: Symbolic Interactionism: Feminist Theory: Postmodernism: Critical Race Theory: Social Exchange Theory: Rational Choice Theory: