The trusted cloud has become the number one priority to better protect our industrial data in France and Europe. But for Antoine Couret, founder of the start-up Aleia, this sovereign project should also contribute to the emergence of a trustworthy artificial intelligence as such. This Monday, he is taking advantage of the opening of the Big Data & IA show to announce the launch of the first version of Aleia, an open, collaborative platform and “sovereign” guarantee.
Why did you launch Aleia two years ago?
Antoine Couret: Aleia was founded at the end of 2020 with the conviction that artificial intelligence (AI) will revolutionize industrial sectors. But to take advantage of this transformation, the industry must have easy access to sovereign cloud and AI. We wanted to provide this trusted AI platform that is not limited to the production of experiments or Proof-of-concept.
In the same way, these industrial processes must connect to a broader ecosystem. Our platform proposes to build not only a free access library of algorithms but also an ecosystem of actors around these issues. Our wish is to enable them to have a common language and to be able to pool their resources. This is also the whole point of an initiative like Hub France IA which is committed to boosting the AI ecosystem to accelerate the development of operational projects.
It must be said that in Europe we have two assets in this area: we have major industrial leaders and cloud providers. But these transformation issues were not as significant two years ago and are now involved in complex issues in terms of climate, energy and even food transition. European governments have realized the need for more industrial sovereignty in response to the recent weakening of energy and logistics chains at the global level.
If we have developed the “IA as a service” platform ofALEIA and that we are launching it commercially today, it is precisely to be able to provide a complete, accessible and immediately deployable base for industrializing AI projects.
Aircall, the centaur of cloud telephony
What are the steps to keep in mind to guarantee the success of an artificial intelligence project?
In my opinion, there are three crucial steps, which we observed during the development of the platform, and which also allowed us to adapt it as closely as possible to the needs of the customers. The company must already have understood the need to transform into a real “data & IA driven” organization. In other words, the fact of including the processing and interpretation of data in its strategic decisions. Most organizations have understood the interest, but a common language is still needed to better integrate the professions.
Many companies have accelerated on the storage of data grouped in datalakes for example, but if no fine algorithm is planned, it will be difficult to take full advantage of their data. For example, on the subject of the carbon and energy balance, a company is led to analyze all its purchases as well as all the data of thousands of different actors distributed in its ecosystem. This can take a long time and she is often forced to resort to AI.
The second thing to keep in mind is the triptych “IA/data, IT and businesses” because there is no point in deploying algorithmic tools if the people concerned do not understand how to use them. And finally, the third piece of advice I can give to companies is to be reasonable. Start with a simple project that has both business and IT benefits. The professions must understand the interest of the AI project and the data scientist should try to understand their needs.
It is therefore advisable to limit this first project to a motivated group that can act as a catalyst that takes the company on board towards a deeper transformation. And also advised to rely on partners who are already experienced to support in the implementation and deployment of the project to go beyond the experimentation stage and find a certain sustainability of the AI approach.
These three steps are at the heart of the collaboration enabled by the platform ALEIA : exchange with the trades, get quick results, ramp up the company on AI.
After wanting to regulate the use of personal data in recent years, the European Commission now seems to be taking a closer interest in industrial data… What do you think?
The awareness around industrial data is clear. Unlike personal data, which has been mostly captured by tech giants, industrial data is still in Europe. Industries understood that they were sitting on a gold mine.
Regulations to better protect this data are welcome, but we will also have to take an interest in the technologies used to process them. The cloud of trust is necessary but not sufficient to embark our industry towards AI. We will have to pool European offers and encourage industries to demonstrate greater maturity in their project experimentation – with the best datasets and the best algorithms.
Some projects have already proven themselves, such as Agdatahub, a platform allowing a better circulation of agricultural data in France and Europe. Others such as IA Cargo or Renovaite also make it possible to optimize logistics or energy chains and reduce overall fuel consumption. In this sense, AI is an effective way for companies to reduce their carbon impact, while saving money.
But to exchange data securely and thus generalize this type of project, we need a trusted platform. And the regulatory framework in this area is welcome.
I am convinced that SMEs and ETIs are the key ecosystem in the evolution of digital transformation. With the major contractors, they are very numerous and have a power of reciprocal irrigation of AI subjects
Aren’t these stricter regulatory frameworks on data also likely to complicate your task?
These regulatory waves are quite broad and the fact of wanting to protect our industrial data in Europe is positive. When China imposes on Apple to give access to all its data stored on its territory, I do not see why Europe could not impose the same thing.
However, we are in a period of transformation that is difficult to define and we must also plan for areas of discovery that are not affected by regulations to allow breakthrough innovation. Europe sometimes tends to want to regulate too precisely and quickly and it is desirable that the control of the execution of the standards be done a posteriori. If it were a priori, it would greatly hamper innovation.
On the occasion of the inauguration in Strasbourg of the brand new datacenter ofOVHcloud on September 12, the Minister of the Economy Bruno le Maire strongly encouraged companies to obtain SecNumCloud certification… Do you agree?
The certification SecNumCloud has real potential for strategic differentiation. But it should not be limited to the cloud part and also integrate the application level. The value is not only in the infrastructure but also in the way the data is used via AI algorithms. The fact is that an industry today knows how to protect its know-how through patents, but when it comes to an algorithm, it is immediately more complex.
During the inauguration, I took the opportunity to express to the new Minister in charge of digital Jean-Noël Barrot the need to extend this label to PaaS, IaaS and SaaS technologies. Especially since the SaaS ecosystem is much richer than that of the cloud: it contains many smaller players and for whom it is very difficult to obtain certification. These companies also do not have the necessary resources to pay two people in their team who would be dedicated to obtaining a label.
So the key lies more in the transformation of small and medium enterprises?
I am convinced that SMEs and ETIs are the key ecosystem in the evolution of digital transformation. With the major contractors, they are very numerous and have a power of reciprocal irrigation of AI subjects.
But today the offers on the market (like those of the hyperscalers) require the user to choose between thousands of components. And too often the language used is far too tech-savvy for smaller businesses.
Aleia greatly simplifies access to AI: our platform is accessible in one click and offers a ready-to-use and very business-oriented environment. At the same time, we are also participating in the Ile de France region’s “IA Pack” program intended to support the first AI projects of SMEs. This year, 200 projects have seen the light of day thanks to this initiative.
In other words, do you want the AI to be much more consumed locally?
This clearly contributes to trust and the emergence of a common language. This proximity gives rise to more interesting research reflections than through hyperscalers. And the more the ecosystem brings together small and diverse players, the richer the emulation.
In the 1950s, aeronautics included around a hundred companies and a few decades later the market refocused around a handful of players. In the AI market, we are a bit in the 50s of aeronautics and we are just beyond the discovery phase.
The objective is not to erect our own hyperscalers but rather to support an organized ecosystem dedicated to the industrialization of AI. Since hyperscalers require the development of strictly siled AI production chains, Europe must promote its culture of openness, allow more interoperability, the basis of this confidence that it knows how to bring.
At the start of the year, you raised 8 million euros… What are these funds used for?
This round table allowed us to recruit around forty people and above all to develop our platform. We are also taking advantage of the Big Data & IA show on Monday to announce the launch of the first version of our sovereign and collaborative “IA as a service” platform which brings the best of AI to all companies.
It is a trusted environment, totally open and dedicated to the production of AI projects. It is the perfect tool for collaboration between teams, between companies wishing to work on different data sets and in complete security.
We would love to say thanks to the writer of this write-up for this amazing web content
Antoine Couret, founder of Aleia: “The cloud of trust is necessary but not sufficient to embark European industry towards AI” – Forbes France
Check out our social media accounts and also other pages related to themhttps://www.ai-magazine.com/related-pages/