Do you want to know if the salmon, hake or any other fish you are about to buy at the market meets the criteria of sustainable development? The Fish Me Well app will tell you. You take a picture of the fish and the artificial intelligence that drives this app tells you if it is allowed to fish in this season, from what size and warns you if it is a female in gestation, the capture of which is normally prohibited.
But don’t look for Fish Me Well on the App Store or Google Play. “This is one of the projects carried out by our students as part of their ‘Understanding AI’ courses at the Makers Lab, our laboratory dedicated to learning by doing”, explains Chloé Pelletier, head of this course at the EM Lyon Business School. Particularity? The students who prototyped Fish Me Well are not data scientists or computer scientists: they did not write a single line of program. To develop their prediction model from photos of fish, they used no-code AI, artificial intelligence software that does not require code.
A tangible economic reality
These platforms are constantly multiplying: the Americans Akkio (for data available in the form of a spreadsheet), Lobe.ai (for image recognition, owned by Microsoft since 2018), Obviously.ai, the Chinese EasyDL (by Baidu ), the French Prevision.io, the Swedish Peltarion… to name but a few. “AI no-code has been around for several years, but has only become a tangible economic reality for two or three years,” notes Florian Douetteau, co-founder and CEO of Dataiku, a French company created in 2013, specialist in the science of the data and which in turn launched into no-code AI.
“The principle is very simple: in theory with an AI no-code tool, a business expert, specialist in marketing, sales, finance, logistics, production… is able to take data from his activity , to import them on one of these platforms, to apply some filters and adjustments and to develop in a few hours a predictive model directly usable in his company: for example, to predict the customers about to change telephone operator”, explains Julien Laugel, chief data scientist (data director) at MFG Labs, the entity dedicated to data processing and data science, of the Ekino consulting group.
Lack of manpower
The current craze for no-code AI is of course part of the more general trend towards low-code/no-code, which makes it possible to develop computer applications by coding minimally, or even not at all, thanks to a graphical interface and drag and drop. But it can also be explained by the bottlenecks that companies, more and more numerous, wishing to get into AI are currently experiencing. “It’s not a problem of data availability,” stresses Florian Douetteau. Thanks to the falling cost of cloud platforms, companies can now centralize 90% of their information, at a reasonable price. »
What is missing is the skilled manpower to mine this material, which creates a vacuum for no-code AI. “Our typical customer only has a small team of data scientists managing many projects,” says Jonathon Reilly, co-founder and COO of Akkio. Therefore, they don’t have time to create applications for sales people, marketing or logistics managers, but with these tools, these managers can manage on their own. “But there is a very big market for AI, judge Khalil Alami, a data scientist who created a start-up, Quarks, which should offer this summer a no-code AI application based on small components, easily assembled together, like Lego. We can imagine tomorrow using an AI to develop a complete marketing campaign, up to the production, in virtual images, of the advertising spot. » Last advantage: all these solutions being online, it is not necessary to install additional computing power in the company.
The editors of no-code AI solutions are of course rave about their algorithms, even if their handling is not always as simple as some claim (see box). “We help businesses and public sector organizations turn data into actionable insights much faster and more accurately than humans could alone,” swears Matthew Zeiler, CEO of intelligence platform Clarifai. to model image, video, text and audio data, with little or no code.
Researchers are more cautious. “By allowing data to be manipulated without programming knowledge, no-code AI has undeniable educational advantages, recognizes Jérémie Sublime, teacher-researcher in data sciences and AI at Isep, an engineering school in the digital. But be careful, in business, not to play with fire: using this type of tool without understanding their internal logic or their limits can make people believe that AI is an infallible oracle and miss out on major societal issues such as possible biases in the data. »
What do users think? “We have tried, without much success, to use no-code Ai on complex problems of optimizing our metallurgical processes, relates Jean-Loup Loyer, chief data officer of the French mining and metallurgical group Eramet. It is an educational tool that makes it possible to get twice as many employees, non-specialists in AI, to adhere to the challenges of this technology, but on less ambitious projects. »
Another limit: the DSI (information systems department) must give its consent before the commissioning of the final application throughout the company. First of all to avoid a “shadow IT” phenomenon, where we no longer know who does what, in terms of IT, within an organization. And to validate an AI which, as a decision-making tool for certain executives, can have a formidable impact on the smooth running of the company. Finally, as with everything that happens in the cloud, the IT department often remains more competent to ensure that the data is stored in France or in Europe…
Advantages but also disadvantages
· A few hours of training are enough for a first introduction.
· The different stages of the data flow (provision in the form of a spreadsheet, cleaning, choice and training of AI models, verification of the model, etc.) appear in a very educational way.
· Similarly, it only takes a few hours to choose a new dataset or new cleaning filters, a new AI model and see the resulting changes in results.
· No explanation is given on the results obtained, which reserves these services for simple business cases and excludes any use for medical purposes, for example.
· Some platforms reserve the right to reuse their customers’ data to improve their performance, which may mean that they can use it to retrain their AI models.
· All these platforms do not provide for a systematic control of their use; some ill-intentioned people could thus divert image recognition to monitor their neighbours…
Tools not so easy to access
We tested the no-code AI platforms of the American Akkio and the French Dataiku, asking them to predict how often the short text that we will publish on LinkedIn to announce the publication of this article in “Les Echos” would be seen on this social network. Good news: the two AIs converge on the same estimate, namely some 1,100 “impressions” (go to my LinkedIn account to check this prediction…). Advantage: these tools present an operational interest, since a small interface makes it possible to enter a new version of the text and to see if the forecasts increase or decrease. Downside: the first use of these tools is not at all obvious and requires at least half a day of training and some basic knowledge of data filtering, error rates, etc. “These are tools that can be tamed, recognizes Eneric Lopez, director in charge of AI at Microsoft France. These pre-trained AI models are aimed at profiles that are between enlightened business experts and data scientists, a bit like when Excel and Business Analytics software began to become democratized. »