IA: Between Data and Automation (RPA)

Share this post

Who hasn’t heard anything about Artificial Intelligence?

For some, it is scary and for others it is quite science fiction. One thing is sure: we are immersed, sometimes even without knowing it, in a world that is increasingly “riddled” with AI. Siri, Alexa, Netflix, our cell phone and so much more. The world around us surrounds us with artificial intelligence sometimes even without our realizing it. But then, what makes us agree to let ourselves be “enslaved” by all this technology at the very heart of our daily life? Is it out of laziness, a technological challenge or quite simply for greater profitability?

The limits of Artificial Intelligence (AI)

One thing is certain, Artificial Intelligence will change (in fact it is already changing it) the way we perceive the world, but can it do it alone? It is clear that our society has already changed in relation to these technologies. But what will be the magnitude of the upheaval? Will there be a Big Bang? Will robots dominate us as James Cameron illustrated with Terminator in 1984?

Without going to such extremes, I think we have to put things into perspective while nevertheless taking the scale of the phenomenon as it is. Certainly Artificial Intelligence is capable of incredible things, but it is still in its infancy and we are very far from a domination of robots (phew)! As such, Yann Lecun further explains that:

Artificial intelligence is not accompanied by a desire for domination, unlike men.

Yann Lecun (founder of Facebook’s AI research laboratory)

Demistifying the AI, he also explains that the AI ​​needs a hundred hours to reach in Atari games the level that a human would reach in 15 minutes! … Sarah Connor can therefore rest easy🙂

Let’s remember two things:

  1. This technological transformation that is the advent of AI in our lives should not be anxiety-provoking. We are just going through an industrial revolution.
  2. AI needs a lot, but then a lot of data to be effective.

I will add to these two points an element that seems essential to me:

An Artificial Intelligence is dedicated to a skill is only one (at least for now).

This is a capital point which also makes it one of its great limits, especially compared to a human.

Breaking the limits of AI

The two main limits of Artificial Intelligence are therefore the quantity of data necessary to create it and its dedication to a task or rather to a precise skill.

Data first

Why then is a lot of data necessary for its creation (modeling)?

Quite simply because there is no intelligence without learning and there is no learning either without an adapted and favorable environment. Artificial Intelligence therefore does have a fuel… and this essential fuel is data. As a corollary, the more complex the intelligence, the more fuel it will need!

That’s the whole problem. When we have to create an artificial intelligence we must first of all ask the question of the dataset that will allow the creation of the underlying model. And if the amount of data is not sufficient, the relevance of the model will not be there.

Big Data will have a real use here .

The quality of Artificial Intelligence is therefore proportional to the quantity of data that will allow its modeling.

But be careful, just like with your food, we must not water our AI with just any data. Otherwise, indigestion is guaranteed. No, it must reflect reality, in its quality, proportion and nature.

The dedication of AI to a skill

We said it beforehand: Artificial Intelligence can be very powerful. But it is only suitable for one task for which it was designed. It’s impossible to ask Netflix’s AI to sort out your musical tastes (while it does for your movies / series).

What then to replace tasks that require multiple skills?

In fact, Artificial Intelligence must be included in the broader sense and included in a more global framework. We often hear about Uses-Cases in the company. What, in fact, is more concrete than business processes that we would like to automate, isn’t it? Unfortunately these Uses-Cases very often (if not always) call on several skills. The only solution will therefore consist in combining Artificial Intelligence in an equally intelligent and orderly manner in order to reconstruct an efficient business process.

Here we are, therefore, we must find the binder of these Artificial Intelligence.

Fortunately, the process automation (or RPA ) approach is not new (either) and is naturally emerging as the backbone of business processes. What could be more logical then than to affix artificial intelligence bricks.

Moreover we can automate simple and basic processes via RPA without intelligence. But it’s so much more interesting to automate or rather to reproduce human behavior in front of the computer. It is here, moreover, that the combination of AI + RPA becomes really exciting.

Isn’t this also the primary objective of Artificial Intelligence: to imitate the behavior of a human being?

Here then is the appearance of the IPA (We call the combination RPA + IA, the IPA as Intelligent Process Automation) which allows to combine Artificial Intelligence with concrete processes.

But then …

If Artificial Intelligence “explodes” and spreads everywhere, it is the same for its entire ecosystem. It is rather logical because if Artificial Intelligence needs data to exist it also needs to find a concrete use, in particular via RPA. Data propagation and RPA / IPA are therefore spreading today in all levels of companies but also at the heart of our lives in order to optimize the tasks that have sometimes become complex in our digital life.

Good or bad thing ? Difficult to say yet so much the road and long. Personally, I would say that this is neither positive nor negative: we are going through a technological revolution, and as with any revolution or big change, we must adapt and tame it.

One thing is certain: our personal and professional lives generate a lot of information / data. This data is immediately reused to get to know us better and thus reproduce our behavior. It is a fact and it is inevitable! It is up to us, human, to know what we want to accept and how we will adapt to this great transformation which has already started.

I already have my idea on the subject … and you?

Share this post

Benoit Cayla

In more than 15 years, I have built-up a solid experience around various integration projects (data & applications). I have, indeed, worked in nine different companies and successively adopted the vision of the service provider, the customer and the software editor. This experience, which made me almost omniscient in my field naturally led me to be involved in large-scale projects around the digitalization of business processes, mainly in such sectors like insurance and finance. Really passionate about AI (Machine Learning, NLP and Deep Learning), I joined Blue Prism in 2019 as a pre-sales solution consultant, where I can combine my subject matter skills with automation to help my customers to automate complex business processes in a more efficient way. In parallel with my professional activity, I run a blog aimed at showing how to understand and analyze data as simply as possible: datacorner.fr Learning, convincing by the arguments and passing on my knowledge could be my caracteristic triptych.

View all posts by Benoit Cayla →

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Fork me on GitHub