Democratizing Artificial Intelligence
The rise of the Personal Artificial Intelligence
We are moving towards a future when every person will have the capacity to train their own Personal Artificial Intelligence system using their own data.
A known path
Every technology travels a similar path, it starts as an experiment, without a clear practical utility; later evolves into a limited access system, usually rented to third parties for their use; finally the technology becomes efficient enough to make its way directly into the consumer’s hands.
We saw this on transportation, with the passage from the train to cars, on the computer industry with the passage from big computers to personal computers, on the media with the passage from television and press to streaming and blogs.
As an example Clementina, the first computer to run in Argentina in May 15 1961 was used to validate the Hally Comet’s orbit, to simulate telephonic traffic and diverse tasks related to the academic and business worlds.
Today we can calculate the trajectory of each known celestial body with our smartphones in real time. But to train a machine learning system we have to go to big companies, in a way not unlike how it happened in Clementine’s time.
Just as access to computer software is practically universal now, the same should happen with different machine learning and artificial intelligence systems.
“Why would an ordinary person want to train a machine learning system?” might be asking more than one reader. The same question arose during the times of the first PCs (and cars, electricity, etc.), and in general the answers were very limited in comparison to the applications that were finally discovered.
Originally, PCs were presented as a tool for home accounting, or as a word processor. However, with the passage of time, and thanks to the creativity of the users, we found an infinite number of different applications.
The same will happen with Personal Artificial Intelligence, once people have direct access to them, an infinite number of applications will arise and they will become an everyday part of our lives.
What applications should we expect?
Ana is the (hypothetical) owner of a handmade ice-cream parlor. In the past, every day she had to calculate how many kilos she had to produce in order not to end with excess or be short. It was a stressful moment for her.
Throughout the year she entered the number of kilos sold day by day in her personal AI. Now she can simply consult the estimate of the demand for the next day or week. The day begins smoothier knowing that she has an estimate to rely on.
Marcos has 3 children 8, 13 and 16 years old; and a very complex routine. He has to remember everyone’s daily schedule and tried to help himself with a calendar. But it was useless, it was impossible to keep it up to date.
Marcos was entering each activity into his personal AI. He now receives proactive notifications that take into account external data such as weather and traffic. If there are any unforeseen changes, the AI will adapt, modifying schedules by delaying or anticipating notifications to maintain the consistency of the daily routine.
The building blocks
In order to reach this point, a series of advances must be made, which in principle are not technically very complex, but nobody has worked on until now.
First of all, we need to develop an open format that allows people to store the data, a system of Event Cognition based Knowledge Representation. Which will function as a Universal Dataset, allowing to store arbitrary data and arbitrary relationships between data.
Secondly we need to work on a simple but detailed way to visualize and edit this data, so that anyone can make changes and add more information. The user experience will be a combination of Voice Interaction with information displayed in a three-dimensional space where Mixed Reality will play a fundamental role.
And finally we must advance on automation in the areas of Artificial Intelligence and Automatic Learning. In such a way that algorithms can adapt quickly to the data provided, without the need for laborious adjustments.
What an exciting time to be alive!
Artificial intelligence is everywhere. All large companies bet that the large amount of data and computing power they can access allows them to create more efficient tools … in manipulating our consumption habits.
I believe that artificial intelligence has not yet been exploited to its full potential to make our lives more efficient in the everyday. In things that may not be important to Google, Amazon, Apple or Facebook, but that are important to us.
Artificial intelligence has to be not only the new electricity, but also the new PC.
- More information about Event Cognition based Knowledge Representation
- You can support my work whishlisting my AI game on Steam.
Eibriel
eibriel.eth.limo
