By Paul Jurcys
If you could ask your personal AI assistant anything about your personal life what would you ask? What would you like to know about yourself based on your own data?
Based on my current level of restfulness, if I want to have a productive day tomorrow, how should I prepare for it? Based on my data about this week’s exercise records and routines, what improvements can I make to increase my performance next week?
The current ChatGPT boom has spawned many new generative and general intelligence startups utilizing different large language models. These models are trained on public data sources and can answer various questions based on this training data set.
However, they have limitations as it comes to answering more personal questions and with privacy concerns, there are opportunities to build new AI models on personal data frameworks like Prifina’s user-held data model. In this post, we explore how personal applications and “AI” solutions can be realized on a more individualized basis.
Can Everyone Train Their Own AI?
We are at the turning point where the human-centric technology ecosystem becomes real and individuals actually control their own data. Federated cloud architecture and edge computing technologies have now matured to empower individuals with their data.
Prifina’s human-centric data model refers to a data infrastructure that is built around an individual consumer who can collect data from various personal data sources in real-time and have actual ownership and control of such data.
This human-centric approach to personal data is built on the principle that an individual’s personal data should be private by default and not shared with anyone unless the user decides otherwise.
The idea of individuals owning their data is not new. In the early days of the internet, entrepreneurs working at up-and-coming companies thought that user’s data should belong to them.
Being able to access, collect and own your personal data from different sources is the major condition for individuals to train their personal AI.
Prifina's User-Held Data Ecosystem
Prifina’s user-held data platform has two distinct features.
First, each individual user has their personal data cloud where they can collect data from various data sources: personal sensors, wearables such as smart watches, fitness or sleep trackers, and online data sources (e.g., the location from Google Maps, Amazon shopping history, activity while using social media services, apps, etc.).
In Prifina’s ecosystem, each personal data cloud has an embedded software robot that unifies the data regardless of the format it comes to the personal data cloud.
Second, developers and companies can use Prifina’s open-source tools to build new applications that “come to the user” and run locally (i.e., “on top of” user-held data).
Instead of sending data to centralized siloed environments of iOS or Android or any other company providing a sensorized device, individuals in Prifina’s user-held data environment can benefit from full ownership and control: personal data is private by default.
Prifina’s user-held data model opens new opportunities with personal data: developers can easily build apps without having to solve complex data (back-end) problems. Prifina’s approach to personal data frees developers from the hefty burden of complying with data privacy regulations because apps run locally (“on top of” user-held data) and generate value from personal data on the user’s side (instead of being centralized in a service provider’s platform).
Understanding Your Data with Personal AI
If you think about your own personal data - data from your personal fitness trackers, smart watches, personal IoT devices at home, and all apps on your phone - are you actually able to benefit from all the data you have generated over the years?
At Prifina, we start from the premise that each individual should have a master copy of their data. In our user-held data ecosystem, individuals can connect all their data sources and create their own personal data hubs.
Once you have all your personal data in one place (your personal data cloud), you can start exploring it: your personal AI assistant will answer questions about your own personal data (not random public data that has no relevance to you).
For example, you can ask your Personal AI assistant how well you slept last night, or after finishing your daily morning exercise, you can ask your personal AI assistant about your performance today and what circumstances during the past few weeks contribute to better or work physical condition.
Here are some more specific examples of Personal AI assistants that We at Prifina built to empower individuals with their own personal data:
Example #1: My personal avatar
At Prifina, we built a personal avatar that takes into account data from an individual's personal fitness devices and, based on various biomarkers, appears differently. So, if you were staying well hydrated, got enough sleep and also managed to exercise, your avatar appears fit and energetic. However, If you worked long hours today and did not have enough rest last night, the avatar will appear sluggish and unhappy.
You can also interact with your personal avatar and ask: “what is my readiness level now?” “How can I improve my productivity today?” The avatar will give you a response based on your own personal, and if relevant, some public data (e.g., weather conditions during your recent activity, etc.).
Example #2: My Data Diary
Imagine if you could connect multiple data sources to your daily calendar: your wearable fitness devices, sleep trackers, home IoT devices and your online accounts (such as Spotify or Netflix, etc.) and that you talk to your data diary, what would you ask?
For instance, you could ask your data diary about your readiness level before and after the meeting, or how focused you were before the meeting.
Based on your wearables and other data, you could also ask your personal data diary about the music or surrounding circumstances that make your day more productive, wouldn’t that be great?
What else would you like to explore with your intelligent data diary?
Empowering Developers: Liberty. Equality. Data.
Prifina’s open developer platform empowers developers to build applications that can correlate different data sources - both private and public.
At Prifina, we are gradually releasing resources for developers to build on top of user-held data and to integrate various machine-learning tools into apps. These tools help developers and brands build applications to help individuals get adily value from their data - to interact with their data and ask anything.
Consider all possible use cases where individuals could benefit from their personal data: health and wellbeing, emotional state, connected home and offices, optimizing daily routines, and so on.
In Prifina’s personal data infrastructure, personal AI tools can be embedded in any application. Such personal AI assistants could take different forms: a simple chat window or voice-to-text where you can ask your personal AI any question and get an answer based on your personal data and publicly available data.
Most importantly, Prifina’s user-held data model ensures the privacy and safety of personal data because every user’s data is private by default.
Welcome to the human-centric data ecosystem, powered by Prifina!