"The AI company within everyone's reach" was the title given to Dataiku's EGG Paris 2019 conference, which took place on 7 November at the Maison de la Mutualité. Forecasting AI had made the short trip to give you its feedback in two acts.
Did you miss it or weren't invited at all? Don't panic, Forecasting AI is back for you on the EGG Paris 2019 conference organised by Dataiku at the beginning of November at the Maison de la Mutualité.
EGG, what is it?
The EGG conference is first and foremost an event organized by Dataiku in Paris, London, New York and San Francisco. In view of the growing interest of populations in Artificial Intelligence in a period, shall we say, of transition between an industrial past and an ecological future that many people are calling for of their own free will, Dataiku naturally thought that there was material to create such an event. EGG is the birth of a new society transformed by new technologies in general and AI in particular, and is intended to be the largest conference dedicated to Corporate AI.
It is also an opportunity for the largest companies such as Microsoft, OVH or La Poste to present their innovations in their AI and data science projects to a select audience (a mix of prospects, collaborators or partner companies).
"The advent of the cloud heralds the explosion of data volumes and the diversification of digital information sources that will help to feed and nourish business strategies, in search of more performance, responsiveness and productivity. With these words, Florian Douetteau, CEO of Dataiku, sums up the importance of holding a conference dedicated to Data Science and the AI company. Faced with these business challenges, the AI plays a key role in facilitating the collaborative work of business, functional and IT teams and thus allowing the inclusion of knowledge and know-how held by all human beings within the company."
It was it that opened the ball of this crazy day (or not) with an opening keynote soberly named "Can we Make AI Likeable". A keynote we won't tell you about! Yes, that's the way we are at Forecasting AI. But, as we feel a hint of bitterness rising inside you, we still explain why. This inaugural keynote is indeed a simple reworking of a speech dated 17 September 2019 already available on Youtube.... Warmed up? We leave you alone as a judge but we will therefore skip this presentation (which you can still see comfortably seated in front of your screen if you wish).
Microsoft and the responsible AI....
In any case, the foundations had been laid, all that remained was for the speakers to enter the contest, to take possession of the main stage of the Maison de la Mutualité to talk about responsible AI, semantic AI or even data governance. A busy, dense, very (too?) dense program with Bernard Ourghanlian, Chief Technology & Security Officer of Microsoft, to succeed Florian Douetteau and definitively launch the day! Everyone was thus entitled to fifteen minutes of speech on topics related to data or AI.
This may sound like a smile, but Microsoft, at least its representative, started by talking about AI and responsibilities. What should we remember from these fifteen minutes? Well, Artificial Intelligence has a societal and technological impact on issues such as employment, wealth distribution and security. That it cannot therefore remain too far removed from ethics for too long to be fair, reliable, safe or inclusive and responsible. Bernard Ourghanlian then explained Microsoft's choice to create an ethics committee called AETHER (AI and Ethics in Engineering and Research) with experts from all walks of life. The objective is simple: to judge the feasibility of the project according to the ethical question. What will be the purpose of the system? Will it have a positive impact on individuals and society? What will be its technological capacity? Will it be of high quality and/or reliable? Finally, will its use be sensitive and have a significant impact on individuals and society? Without forgetting to mention that all Microsoft employees are aware of ethical issues, Bernard Ourghanlian therefore points out that several projects have already been retooled by the ethics committee set up by Microsoft.
La Poste, the transformation is now
It was then the turn of Pierre-Etienne Bardin, Chief Data Officer & Head of Analytics, to enter the "arena" to discuss the place of data within the La Poste group. On the agenda, a 4-step attack plan to become a data-driven company without having to create a datalake from scratch. Four steps, detailed by the CDO, were necessary for this transformation by highlighting the creation of value before scaling up.
To do this, it was first necessary to raise awareness of AI, ethics or data and its charter through e-learning or awareness sessions. A first step that Pierre-Etienne Bardin calls the "Data D of DNA" and which precedes the "Data as an asset". It is a matter of mastering your data, its governance or even having DGMP and other compliance certifications. Finally, La Poste ultimately created a Public Data Lake (reserved only for La Poste employees but democratic) before moving on to use cases based on AI to improve ROI and motivate business experts.
Later in the morning, OVHcloud, ENGIE, UNICEF, Levi's and BNP Paribas took over on the main stage to discuss the transformation of their activities linked to the emergence of data and AI. The afternoon was a la carte with more specific and longer workshops (30 minutes against 15) depending on the room chosen among the three available. You could then hear about the choice of Data SEO with the training company of the same name, Artificial Intelligence and data to clean the oceans with The Ocean Clean Up or the state of the art and best practices for AI with Dataiku under python. Use cases that we will detail during our second meeting on the subject.