What is lifeline?

Life Line is a reliable emergency response app designed to provide quick and effective assistance during critical situations. The app uses Google's Gemini API to provide users with AI assisted emergency response, aid and Diagnosis through an experimental AI doctor feature.

Inspiration

The medical care system in a developing nation like India is severely overburdened, with a doctor-to-patient ratio far below the World Health Organisation's recommended standard. Emergency response times are often slow, particularly in rural areas, where the average response time can exceed 30 minutes. Doctors are over-stressed, frequently dealing with an overwhelming influx of patients that prevents them from giving each patient the proper time and attention needed for effective care. According to a 2020 report, India had just one doctor for every 1,457 people, well below the WHO's recommendation of one doctor per 1,000 people.

I have personally experienced the shortcomings of this system. In a critical situation, the delayed emergency response and the rushed consultations made it evident that something needed to change. This experience inspired me to build Life Line, a solution aimed at improving emergency response times and providing preliminary medical assistance to alleviate the pressure on healthcare professionals. With AI based emergency dispatcher and an AI Doctor for diagnosis and preliminary assessment, Lifeline is a complete solution to provide healthcare to those who cant access it.

What it does

  • Quick Access to Emergency Services: Connects users to emergency services instantly using Gemini API.
  • AI Assisted Diagnosis: Experimental AI doctor feature revolutionise healthcare by providing accurate and quick diagnosis to patients who may not have access to proper healthcare.
  • Regional Language Voice Input: All these features can be used using voice commands and inputs. And these inputs are available in regional languages (Hindi) for easier access of masses.

How we built it

  • The app was built using Google Gemini API. We used React, Redux and Tailwind to build this app.
  • The prompts for the ai model was designed such that it doesn't trigger the safety flags of Gemini, which prevents it from doing diagnosis or other such activities which can make it "liable"

Challenges we ran into

  • The Gemini API was not designed for making diagnosis. So the prompts had to be designed such that it could bypass the safety flags and provide diagnosis for experiment purposes.

Accomplishments that we're proud of

  • The diagnosis of this app and its capabilities have been tested and verified by real world doctors and they have agreed that the idea does have potential.
  • The app has successfully diagnosed complex diseases such as Pneumonia, Meningitis, Abdominal Aortic Aneurysm, Parkinson's Disease, and many many more.
  • The ai can study medical imaging scans and reports and give accurate diagnosis.
  • The Emergency assistance feature can provide rapid response and teach users to do life saving procedures such as CPR.
  • The fact that technology has developed so far ahead that we can build AI-based doctors is incredible.

What we learned

  • An AI-model can do virtually anything, only thing that limits us how we design the prompt in order to "persuade" it to do so.

What's next for Lifeline.ai

  • The future for lifeline holds tremendous potentials. In future, we may have doctors and ai-models like lifeline working together to solve medical conditions.
  • AI will work as a junior doctor, doing preliminary diagnosis and providing referrals in case of serious diseases.

Built With

Share this project:

Updates

Submission history