AI and Infection Prevention (2024) is a new project by Anna Dumitriu and Alex May in collaboration with Dr James Price (Brighton and Sussex Medical School), Dr Sid Mookerjee (University Hospitals Sussex) and Dr Ashleigh Myall (NEX.Q). The project artistically explores novel new research which offers to improve infection prevention and control (IPC) in hospital settings using artificial intelligence. The artists are exploring how the NEX.Q AI infection prevention system works by applying it to an AI expanded historic data set based on John Snow’s research on the 1854 Cholera outbreak in Broad Street, London, considered to be the first epidemiological study ever conducted.
The artists are in the process of fabricating a detailed data set using AI tools, and simulations that bring the Broad Street outbreak to life to see if modern AI could have helped stop the outbreak sooner, or even prevented it from ever happening. The artists are in the process of fabricating a far more detailed data set using AI tools, and simulations that bring the outbreak to life and will show us if modern AI could have helped stop the outbreak sooner, or even prevented it from ever happening. Other elements of the project explore how to engage audiences in understanding the meaning and implications of healthcare associated infection (HCAI).
The portraits are based on historic data from Snow’s book “On the Mode of Communication of Cholera”, expanded on using Chat GPT 4.0 to create an image prompt in Midjourney and upscaled using Topaz. Some of the names are real, most are lost and have been created using Chat GPT 4.0.
John Snow’s map of Soho in London, focussed around the Broad Street Pump, is in the process of being recreated by the artists and virtual inhabitants go about their daily business, including visits to the pump to collect water. The simulation records when and where they have been to form part of dataset for processing by NEX.Q and will soon be scaled up to cover the entire area, and population, covered by Snow’s investigations.
The artists have also generated a small 3D printed sculpture using Meshy AI, a 3D modelling tool that uses artificial intelligence to create 3D models from text or images. This model was generated from the image of Clara Thompson exhibited nearby, using its Extracts Prompts from Images feature. Meshy AI extracted the following prompt: The image depicts a cobblestone street with irregular, rounded shapes in a wet, smooth texture under cloudy skies, creating a gritty, historical feel.Strangely the AI extracted the background of the image, the cobbled street, and missed the central figure. A big challenge in data analysis is to work out what is the salient data and what is just background noise.
Other elements of the project explore how to engage audiences in understanding the meaning and implications of healthcare associated infection (HCAI), how patient data can be accessed and employed in a beneficial way, and public understanding of AI, machine learning, and prediction.
Supported by Brighton and Sussex Medical School and University of Sussex Higher Education Innovation Funds (HEIF), supporting a wider program of work in Sussex on translational clinical research in IPC through collaborations with academia, clinical, industry and patients/public.