Exclusive Interview with Tom Morisse, Research Manager at FABERNOVEL

 

The buzz surrounding AI is huge. However, for many people, Artificial Intelligence is still unknown, while being fascinating and sometimes threatening. Is it going to destroy jobs? What are the ethical stakes? Would it be the magic solution to our problems, from bad diseases to global warming? From Netflix’s impressive recommendation engine to vocal assistants such as Siri or Alexa, we find it in numerous tools we use on a daily basis. And this is only the beginning!

To talk about this burning topic, FABERNOVEL INSTITUTE has partnered with Coorpacademy to release a course called “Introduction to Artificial Intelligence”, to understand it better and learn how the leverage it. On this occasion, we sat down with Tom Morisse, Research Manager at FABERNOVEL.

AI can be a very technical topic and sometimes seems far from daily tasks or daily assignments at work. Then, what’s the use of training your company’s coworkers to it?

 

Now that AI came out of research labs and becomes more mainstream, it’s a good moment to start learning on it. Like for all digital tools, it’s crucial to train their coworkers so they can take advantage and use the several opportunities AI can create. Either to offer new experiences to their customers – for example with very personalized services. Or to facilitate the work on the workplace – for example with solutions which anticipate machines failure on a production chain.

Are AI projects accessible to all companies?

The current dynamic of AI impresses by the obtained results, but more also by their extremely fast democratization. Cloud offerings from the giant companies in this field – Google, Amazon, Microsoft – or from specialized startups are easily reusable by companies from all sizes in their own services.

It is true that to develop a very specific algorithm able to solve a problem, to make it tailor-made, the question of getting access to the right talented people can be an obstacle. But even in this case, the existence of several technological assets in open source keeps democratizing AI.

Which project using AI was the most memorable for you?

The project “Teaching Machines to Draw” from Google was the most memorable… for its poetry. They trained algorithms able to recognize animals drawing, such as cats for example. And even able to redraw those animals from our sketches.

If you draw a cat with 5 legs, the algorithm will draw again a cat with 4 legs – as proof that it understood a cat was normally made with 4 legs. You might say that it’s not very interesting so far… Where it becomes interesting, it’s when you give to the algorithm a drawing that has nothing to do with a cat. Engineers behind the project drew chairs, which resulted into surprising hybridizations.

In this example, we can imagine a designer using the algorithm to generate exploratory ideas, to go deeper in one. This project is then for me the perfect illustration of what technologies can do for us: to help us unleashing our creativity.