Lifeline Pitch - UCL First Response in a Box Design Hackathon 2023
0Lifeline Pitch - UCL First Response in a Box Design Hackathon 2023
Each year, more than 350 hundred natural disasters occur around the globe {source: BA Ferrier, JT Spicke3 “Natural Disasters in Developing Countries: An environmental health perspec@ve” from the School of Public Health, Cur@n University of Technology, Perth, Australia}. And both the magnitude and frequency of these are escalating as climate change becomes a growing concern and threat {Source: The World Meteorological Organisa@on’s report “Weather-related disasters increase over past 50 years, causing more damage but fewer deaths”}, from 39 incidents in 1960 to 396 in 2019 {Source: The institute for Economics and Peace}.
While the effects of these are global, the most impacts are most heavily felt in developing countries, where 91% of deaths occur {Source: WMO’s report men@oned above}, the cause of this is complicated, but in this project we intend to target a particular aspect of this problem in a low-tech, cost-effective and scaleable way, the human cost of disasters and its knock on effects on the long term development of sustainable systems. {Source: BA Ferrier and JT Spicke3’s previously mentioned paper}.
A recovering community cannot focus on the development and creation of resilient systems, and so by decreasing the fatality and maiming rate, the recovery process is accelerated, there will be less people who are unable to work and need benefits or are in poverty whether that be because they are crippled or the primary revenue generator in the family has passed, and therefore there are less people dependent on the government, putting less strain on recovering economies.
We can mobilise a larger, more effective population of volunteers by providing real time, speech-to-speech, interactive emergency aid assistance through the use of Generative AI to provide context based, clear and summarised instructions from information looked up a downloadable database of common medical instructions {proof of concept: summarisation of info can be shown from chatGPT’s ability to summarise papers, documents and essays and database search up is possible as shown by AI powered search engines as well as things like using LangChain with an SQL Database Agent.
This effectively lowers the barrier of entry, increasing the population of viable workforce, without massively increasing mishandling related accidents, maintaining the effectiveness of rescue efforts. This approach has two aspects, depending on the connection of the device: When the device has a stable connection to the public cloud, it fully performs all functions described above, giving speech to speech instructions to enable volunteers to act individually by not occupying hands.
When not connected, it instead focuses on providing a smaller range of emergency first aid services such as Wound Analysis from Smartphone photographs {Source and proof of concept: the paper “Proxy – An Innovative AI Monitored First Aid Bot” by Nalinipiriya G, Janakirim J, Kumar R, Vishvaja J}, whose proper and quick dressing is crucial in preventing chronic wounds { Source: https://www.notion.so/Research-data- 24df674dc8cc49118c8cf4c4243a8fd9?pvs=4#10cb1fab6205420886fdd0e1de9a9dc5} as well as continuing to provide a medical manual with speech to speech support. The availability of both a low and high tech option means this is highly scalable.