Artificial intelligence (AI) is undoubtedly one of the most transformative innovations of our time. It permeates every aspect of our lives, from how we consume information to how businesses operate. Warren Buffett, one of the world’s most successful investors, acknowledged the magnitude of this revolution in 2022, stating that AI will change our society more than anything we have seen so far. However, this is not without raising concerns. The renowned physicist Stephen Hawking, back in 2014, had already warned against the risks of AI. He described it as the greatest event in human history, which could also be the last if we fail to manage its threats.
The etymology of the terms “intelligence” and “artificial” invites us to explore the very foundations of what we mean by “artificial intelligence.” The word intelligence comes from the Latin intellegere, which combines inter (between) and legere (to choose, to gather). It therefore refers to the ability to discern and connect elements to arrive at understanding, synthesis, or decision. This aptitude is one of the most distinctive traits of the human mind. It highlights faculties such as logic, creativity, and even intuition.
On the other hand, the term artificial comes from the Latin artificialis, derived from ars (art) and facere (to make). It designates what is the product of human work or creation, often imitating nature without being an integral part of it. The idea of artificiality emphasizes the notion of imitation and fabrication, a creation that, while shaped by humans, tends to reproduce natural properties.
Artificial intelligence, through its very name, carries an intrinsic tension between the natural and the artificial. It aims to recreate, through technical systems, an aptitude that we traditionally associate with human thought. This capability relies on algorithms, models, and data, which enable machines to simulate understanding, learning, and problem-solving.
The very etymology of the expression “artificial intelligence” reveals its ambition: a bridge between human creativity and technological ingenuity, seeking to reproduce — or perhaps surpass — some of the most complex functions of the human mind. But let’s see how it all began…
History of AI: From Its Beginnings to Current Rise
The history of artificial intelligence (AI) begins well before the arrival of modern computers. In the 1940s, work in computer science and mathematics, notably that of Alan Turing, laid the foundations for thinking about creating machines capable of simulating human thought. In 1950, Turing proposed the “Turing test,” a criterion for evaluating whether a machine can produce intelligent behavior similar to that of a human.
The term “artificial intelligence” appeared in 1956, at a conference at Dartmouth College in the United States. This event brought together researchers such as John McCarthy, Marvin Minsky, Claude Shannon, and Herbert Simon, who explored the idea of building systems capable of reasoning, learning, and solving problems.
The first decades saw the emergence of algorithms capable of solving complex problems and playing chess, for example. In 1965, Joseph Weizenbaum designed ELIZA, software that simulates simple conversation, paving the way for human-machine interactions. However, these advances were followed by periods of slowdown, called “AI winters,” due to technological limitations and overly high expectations. Funding decreased, slowing research.
In the 1980s, “expert systems” revitalized the field. These programs, based on defined rules, replicated human decisions in specific contexts, such as medicine or finance. While effective in specific cases, they struggled to adapt to new situations.
In 1997, Deep Blue, an IBM supercomputer, defeated world chess champion Garry Kasparov. This event illustrated the ability of machines to master complex tasks. However, AI remained limited to specific performances, without reproducing human flexibility.
Starting in the 2010s, machine learning and deep learning enabled significant advances. Based on artificial neural networks, these methods made machines capable of learning from large amounts of data. They found applications in many fields, such as image recognition, automatic translation, and autonomous vehicles. In 2016, DeepMind’s AlphaGo surpassed the best Go players, a game considered more complex than chess.
But What Is an Algorithm?
Imagine it as a conductor directing the operations of a company. It is a set of precise instructions that guide operations and decisions, enabling task automation, information analysis, and solving complex problems.
Take the example of an online store that uses a recommendation algorithm. It studies consumers’ purchase history, preferences, and interactions on the site to suggest products that might interest them. Not only does this improve the customer experience, but it also increases sales and loyalty.
There are two types: first, explicit algorithms, like a cooking recipe, they follow predefined and transparent rules.
Implicit algorithms, on the other hand, learn from data, adapting and evolving over time.
Data: The Fuel of Algorithms
If algorithms are the engine of AI, data is its fuel. It powers them, enabling them to learn, improve, and make intelligent decisions. In the case of an algorithm that predicts weather, it needs data: temperature, atmospheric pressure, wind speed… The more information it has, the more accurate its predictions will be.
There are two types: structured data, organized logically and easily exploitable, such as customer information stored in a database. And then there is unstructured data, more difficult to exploit. Examples include customer comments on social media or emails. Text analysis and natural language processing techniques are used to extract the right information.
If the data is poor quality, it can lead to erroneous results and ineffective decisions. Implementing data collection, cleaning, and validation processes is essential. They are therefore intimately linked. Algorithms need data to function, and data needs algorithms to be exploited.
LLMs: Engines of Generative AI
The term Large Language Models (LLMs) designates a specific category of artificial intelligence models specialized in processing and generating natural language. These models, which take their name from their impressive size, are based on architectures containing billions, even trillions of parameters. The concept emerged with the rise of deep learning and neural networks, but it truly established itself in recent years thanks to major technological advances. Popularized in the field of generative AI, the term refers to systems capable of producing texts, understanding linguistic context, and adapting to various tasks, thus transforming how we interact with technology.
They are trained on massive amounts of data, which allows them to learn the nuances of language and produce coherent and relevant texts. The development of LLMs has accelerated in recent years, with high-performing models generating increasingly sophisticated and innovative texts.
Today, there are more than 450 LLMs, each with its own characteristics. Among the most popular are ChatGPT by OpenAI, Gemini by Google, Llama by Meta, Claude by Anthropic, and Copilot by Microsoft. They differ in cost, accessibility, the number of tokens or text units they can process, and the languages they support. Some, like GPT, can process tens of thousands of tokens, enabling them to generate longer and more complex texts. Other models specialize in certain languages or certain types of tasks.
Use of AI by Digital Professionals
In France, digital professionals use AI to automate tasks, improve productivity, and create new content. According to a recent study, the most commonly used AIs are Gemini by Google and Copilot by Microsoft for text AI, and Adobe Firefly for image AI. ChatGPT by OpenAI is also used by a significant portion of these professionals. These tools enable them to generate marketing content, translate documents, write blog articles, design images and illustrations, and automate repetitive tasks.
The Global AI Market: Exponential Growth
One of the main reasons for AI’s meteoric rise is the explosion of data. In 2023, the amount of data created each day is simply astronomical: 2.5 quintillion. To put this in perspective, imagine that every grain of sand on Earth represents one byte of data. 2.5 quintillion bytes is the equivalent of all the sand on all the beaches and all the deserts on the planet, multiplied by several thousand! Every day, we generate 650 million Tweets/X posts, perform 8.5 billion Google searches, and send 361 billion emails. All this constitutes the fuel that powers the development and learning of AI systems. The more there is, the more AI algorithms can assimilate and improve.
The global AI market is booming, driven by this data explosion and constant technological advances. In 2022, its value was estimated at $177 billion, which already represents a considerable sum. But this is just the beginning. According to forecasts, this market should reach $2,745 billion by 2032, with an average annual growth rate of 36.8%. To better understand the magnitude of this growth, imagine an investment that would double in value every two years. It is dominated by software, which represents the largest share, followed by hardware and services. Companies worldwide are investing massively in AI, aware of its potential to transform their operations and create new opportunities.
NVIDIA is today’s global leader in artificial intelligence. Initially specialized in graphics processing units (GPUs) for video games, it quickly identified their potential for AI and invested in suitable architectures. GPUs, thanks to their ability to perform parallel calculations, have become essential for processing the vast amounts of data required by AI. NVIDIA is an essential player in the AI ecosystem. In November 2024, its market capitalization is estimated at approximately $3,616 billion US dollars, making it the world’s largest company in terms of market value.
Ready to use this bridge between human creativity and technological ingenuity? Will you be an augmented entrepreneur?
