SAFR
Inspiration
The incredibly fast technological development and societal changes have brought immense changes to the way we as humans live. Since the 20th century more and more people have had the opportunity to work and live far away from their hometowns leaving their friends and most importantly their families and parents behind.
Contrary to previous centuries when it was usual for multiple generations to live under the same roof, today many don't even live in the same town as their parents, making it hard despite their best effort to assist them.
The situation becomes especially hard if the individual loses a partner or has no children. The stress of having an elderly family member that lives alone is something many of us are familiar with. Because of current societal changes (falling birthrates, more people living alone, etc.) it's obvious that the number of elderly living alone will increase thus making the issue even more serious.
By assisting the older generation with general health maintenance and helping them in cases of accidents we try solving the main part of this problem.
What it does
We propose a dedicated device (which we simulated), that has the capability of actively encouraging the elderly to note their health related data in a way that is familiar to them.
The basic working principle is as follows: the device calls the owner 1 - 3 times a day asking about their health and general well-being. If the call isn't accepted or declined for 2 - 3 minutes the device notifies the emergency contact who can call the device's owner on phone or decide to notify emergency services.
The integration of artificial intelligence brings many benefits.
There is no need for complex note-taking about health related
information with the help of LLMs natural speech can be compressed into useful data.Older generations are more likely to accept help from a life-like friendly voice.
Chatting at least a couple of minutes a day helps battle the ever emerging loneliness.
The data collected by the device is then saved on a server and can be reached from a web browser. The user can check their health reports going back every day and their data is visualized on charts to give more insight. This can especially be useful for doctors tying together symptoms to make a diagnosis easier.
How does the actual user experience look like?
Using the device:
The device starts ringing and the user accepts the "call", a friendly voice greets the user and asks how > they are feeling, the users answers and the device follows up with an appropriate answer and question > this repeats until the user hangs up the "call" or says goodbye. The second ringing of the day is > declined by the user so no data for that time is collected.
Using the website:
The user has the option to select a day from the calendar and take a look at that days recorded data or > the user can look at the data visualizations provided to gain more insight.
What makes our idea so much more useful than current solutions?
The "Gondosóra" program in Hungary seems to accomplish a similar goal in part, but makes the mistake of not counting with the personal qualities of many grandmas and grandpas that many times feel like accepting and using one of these devices makes them look as someone weak or as someone frightened to live thus making them use the device becomes a source of family conflict.
The SAFR device on the other hand can be given as an easy to use health diary and it's usefulness can be shown off in the online "Health calendar". The fact that the device requires the usual "calls" to either be accepted or rejected, basically the device must be interacted with to avoid notifying family members makes it even more likely to stay in use.
Our device also differs in it being useful for long-term health and having insight into patterns that can be relevant for diagnoses and personal development.
How we built it
We used Python with the Tkinter library to visualize what our device would function and look like. The simulated device summarizes the transcript into a Json file which is saved and transferred to a server using MySQL. The database is read out using Php and is used on our website that shows visualizes and makes the data browsable in an easy way. For speech recognition we use the Microsoft Azure Speech service and we use the LLM ChatGpt-4o-mini vie Microsoft Azure.
What's next for SAFR
There is still a lot of potential in our project, first and most obvious to implement the device in real life, secondly to do more processing on data gathered from the user over a longer period of time, helping uncover hidden patterns and understand more about the users health.
Log in or sign up for Devpost to join the conversation.