AI, what’s it for?


Michel Rubino heads up Cartesium, a start-up software publisher that specialises in embedded artificial intelligence. Here, he gives his expert opinion to help us understand the concept of artificial intelligence (AI) and its impact on society.

“Defining AI is not easy, the word “intelligence” is not really the best choice! It can be said that it brings together different algorithms, which are mathematical formulas that obey specific rules. This set of algorithms will enable software to learn by itself and interpret the environment around it. This is known as machine learning.

By incorporating data (images, sound, vibrations etc), the machine can compare its existing knowledge to new information that it receives and determine differences or similarities. A classic example of this is image recognition software.

At Cartesiam (opens in a new window), we offer embedded intelligence systems: technology that collects and analyses data coming from as close to the physical object as possible thanks to AI and machine learning.

Until now, data collected by users was usually sent to the cloud for analysing it in “cloud computing” mode. Not only is this expensive, only around 1% of the processed data can actually be used. With a product such as our Bob Assistant, the data is directly analysed by the sensor and only important information is sent in which means it’s done at or near the source of the data.

Advances in artificial intelligence are stimulating everyone’s imagination but it can be confusing. Let’s be clear: the artificial neural networks that make up today’s AI are still very far from the human brain! In this respect, it’s necessary to distinguish weak AI – what we know today – to strong AI, which will be able to connect thousands of artificial neurons in the future. This won’t start happening for several decades…

However, one thing is clear: AI will transform the world around us. By 2035, some studies predict it will increase productivity by 40%! This will destroy some jobs and create new ones that aren’t yet known, like in the early days of the Internet.

Whole sections of human and economic activity will evolve. The industry will see widespread interconnection of intelligent machines and objects, and a proliferation of automated processes. In terms of healthcare, AI can already suggest a medical diagnosis based on the patient’s data. As for lawyers, they’ll be able to compare their legal case with the rulings on previous cases in an instant.

The one thing that humans can’t be replaced for is their empathy: breaking bad medical news, helping someone in need, advising a buyer… all this requires emotional intelligence that only humans have mastered.

Freed from repetitive menial tasks thanks to AI, people can concentrate on higher value tasks that could be described as nobler or more beneficial for everyone. This will then prompt the question about how our societies are organised, so that everyone can make the most of this revolution.”

“Artificial neural networks are still very far from the human brain.”
Michel Rubino
Cartesiam, CEO
> Find out more about Cartesiam