Artificial intelligence is certainly the most important technological advancement of this 21st century, although it emerged in the early 1950s, along with the Turing experiments and the first use of machine learning.
Like many technical breakthroughs-industrial machines, transistors, the Internet or mobile computing-AI was thought through and designed academically before being sold, in successive stages. Already present in the daily lives of many of us, however, it offers many additional fields of application to explore.
The first companies to use AI were large technology companies with the scientific knowledge and computing resources to use and adapt neural networks to the demands of their customers. They did this by using the cloud, and scaling AI from academia to real-world business applications.
Google, for example, then applied in-depth analysis of natural language processing to design Google Translate, while Facebook uses AI to identify consumer goods from images.
Tech giants are succeeding in creating platforms that allow companies and startups to benefit from AI in the cloud, faster and at lower cost. Large organizations in all sectors are the first to use this technology more widely to improve the quality, security and efficiency of their workflows.
Data scientists in finance, healthcare, environmental services, retail, and entertainment are beginning to practice neural networks in their own data centers or in the cloud.
AI chat bots support call center teams by testing requests and filtering emergencies. Meanwhile, fraud detection AI monitors unusual activity in online marketplaces. Computer visualization acts as a virtual assistant for mechanics, doctors and pilots, giving them accurate information and making their decisions more reliable, both in real time and in simulation.
Despite the democratization of AI applications (machine-learning, deep-learning, NLP), only 10 to 15% of companies in the world have succeeded in the industrialization of AI-based solutions in their company by 30 to 40%. of these are limited for experiments, according to a study conducted by the company EY*.
In addition to the cloud or data centers, the advent of 5G is in the process of becoming a catalyst that accelerates the deployment of peripheral computing devices, closer to end users: in factories, hospitals, airports, stores, restaurants and within electrical networks.
With the use of IoT devices and advances in computing infrastructure, the proliferation of data is now allowing enterprises to create and practice AI models to deploy “on the edge”, to serve end users and the general public.
AI “at the edge”, already in use in many sectors, appears to be a promising answer to the need to grow very fast and very fast. Computer visualization of those entering and going inside factories detects security violations, checks medical images for abnormalities, and suggests driving vehicles safely on the highway.
The potential for new applications is limitless. Recent advances in technology have even paved the way for a form of AI autonomy, that is, the ability to direct mobile machines without human intervention. Cars, trucks, boats, planes, drones and other robots will soon operate without human pilots.
Over the past 75 years, AI has moved from laboratories to living rooms, improving business performance and the daily lives of consumers. If companies invest in the field, many challenges and barriers to adoption remain.
In many sectors, simulation capabilities will be decisive for modeling and testing in all possible scenarios, and knowing how to predict the security risks associated with deploying robots in the real world. Clear and appropriate regulations are also one of the key elements that have conditioned the emergence of technologies in the daily lives of consumers.
It is only by moving from experimenting companies to well-structured marketing of a product designed or working with AI that its adoption can disappear and fears about its biases can be eliminated.
If many projects remain, it is safe to bet that the new generation of innovators will begin a new era that is synonymous with greater confidence in AI. Not surprisingly, the market for trusted AI is now estimated at 53 billion euros, covering all critical business sectors (automotive, bancassurance, industry, pharmaceuticals, etc.).
* Ernst & Young study for Investment Secretariat