Podcast: our vision of artificial intelligence, by Yahya El Mir, co-founder of iRevolution
Artificial intelligence is a trigger for disruption
Contrary to popular belief, artificial intelligence is not a recent technology. AI is a subject that dates back to the 1950s, and many companies have been working with it for a long time. Indeed, digital has been used for tasks that can be easily modeled, before gradually being applied to human problems, which are much less so. If artificial intelligence has been around for so long, why is it so popular now? There are two reasons for this:
Firstly, it is the result of a revolution that will be at least as disruptive as the invention of the Internet
The Internet also existed for a very long time before all and sundry favored it. The Web browser, and later ADSL, have allowed everyone to connect from home, although initially, the Internet had a purely military or scientific use. The emergence of social networks and other tools in the workplace (email, cloud computing, online backups, etc.) and mobility (smartphones, tablets, etc.) have made the Internet essential to all users. The same stands true for artificial intelligence, the usage of which is only nascent despite its respectable age.
Secondly, we are still witnessing the evolution of the two necessary ingredients for the explosion of artificial intelligence
On the one hand, data collection and intrinsic data accuracy: we have never collected so much data before. The more information there is the more advanced AI will be. On the other hand, computing power continues to evolve. Despite a potential slowdown in Moore’s law, according to which the performance of the best processors on the market doubles every two years thanks to technological progress, this increase in power makes it possible to process all of the available data.
What is the purpose of artificial intelligence?
Artificial intelligence is a pragmatic approach that focuses on solving problems. Compared to computer science, as we knew it, there is no more need to program what the machine is supposed to do by giving it precise instructions. With artificial intelligence, the process is different: the machine will take the execution path.
This technology opens up multiple avenues. Let us take the example of healthcare.
Artificial intelligence offers an opportunity to change the situation in this industry. For example, modern diagnostics are carried out by artificial intelligence that combines data from medical imaging with databases packed with historical diagnoses.
Thus, AI was instrumental in diagnosing dementia, shortening the detection by six years; similarly, for identifying breast cancer five years before it appeared. Finally, in developing countries where access to care is sometimes rudimentary, AI can detect cervical cancer via a simple photo taken by a smartphone compared to an existing database. The results are reliable in 91% of cases, compared to 70% for a conventional diagnosis.
The results of these examples are staggering. The ability for AI to make reliable diagnoses would not have been easily predicted even a few years back. Let alone its ability to perform it better than highly qualified humans. The healthcare sector is on the verge of upheaval, as will other industries shortly.
How to approach artificial intelligence?
With the emergence of Big Data, one heard questions like: “I have a lot of data, what kind of goals can I achieve with this?”. It is time to rephrase this as follows: “I want to achieve these goals, what data do I need to do this?” It is a different approach as it consists in focusing on the purpose, the “why” dimension that explains a real business orientation.
In this business approach, it is necessary to learn quickly. Taking three years to test an idea is far too long a learning curve. On the other hand, taking less than five months to test it and validate this concept in terms of business can make a difference. Proceeding that way makes it possible to launch another larger project to collect all the key indicators, before deploying the idea throughout the entire company.
It is, therefore, necessary to return to basics: technology was made to help you achieve your goals, it is not an end in itself.