Artificial Intelligence was not born yesterday. From the Dartmouth conference in 1956 to today's large language models, AI has been through several cycles of hope and disillusionment before becoming a practical tool for industry. Mimorian is an industrial intelligence platform that models equipment, structures failure diagnosis and captures the know-how of maintenance teams through a multi-agent AI architecture, building on these historical advances.
Understanding this history means better grasping why AI is now present in our factories and in our daily lives. According to McKinsey (2024), 78% of organisations now use AI in at least one business function, up from 55% a year earlier.
How was AI born in 1956?
Dartmouth, summer 1956. A handful of researchers, McCarthy, Minsky and Shannon, came together. Their ambition? Nothing less than to formulate intelligence so precisely that a machine could simulate it. The utopia was set out.
Why did AI experience its first failures in the 1960s and 1970s?
The first steps were promising: symbolic logic, perceptrons. Yet the limits soon appeared. The famous Lighthill Report judged AI disappointing and useless. Funding collapsed. This was the first AI winter. A generation of researchers lost heart.
What did the expert systems of the 1980s bring?
The flame caught again. Engineers coded human expertise into rules. "Expert systems" found their place in industry. Yet once again, the momentum ran out. Too rigid, too costly. A second winter.
How did data revive AI in the 1990s and 2000s?
Computers grew faster, databases grew larger. Researchers changed course: statistics, machine learning. A more pragmatic approach, ready to exploit the growing mass of data.
Why does 2012 mark the rebirth of deep learning?
The decisive moment: AlexNet shattered the image recognition records with a top-5 error rate of 15.3%, compared with 26.2% for the runner-up, an improvement of 10.8 points (Krizhevsky et al., 2012). Thanks to GPUs, deep learning took hold. The world rediscovered AI, this time armed with computing power and massive data.
What did AlphaGo's victory change in 2016?
When a machine beat the champion of the game of Go, a game reputed too complex for computers, the world understood: AI no longer merely follows rules. It learns, innovates and surprises.
How did Transformers open the era of generative AI?
With the arrival of Transformers (Vaswani et al., 2017), a new era began. Language, until then so difficult for machines, became accessible. This was the royal road to ChatGPT, BERT and the large language models.
Where does industrial AI stand today?
From predictive maintenance to industrial copilots, from image generation to report writing, AI is everywhere. The global AI market is estimated at 244 billion dollars in 2025 (Statista, 2025). Platforms such as Mimorian apply these advances in practical terms: a functional digital twin models every piece of equipment, specialised AI agents orchestrate the diagnostic reasoning, and field know-how is captured automatically with every intervention. The architectural distinction between a general-purpose conversational chatbot and a true industrial multi-agent architecture is detailed in our complete guide to ChatGPT versus multi-agent AI in industrial maintenance. The practical application of this architecture to field diagnosis is explained in our complete guide to AI-guided diagnosis in industrial maintenance. Yet behind every advance there were passionate women and men, sometimes overlooked, who believed in an intuition against all odds.
Why it matters for us, industrial players
The history of AI is not linear: it alternates between hopes and disappointments, but always comes back stronger. It is this perseverance that makes it today a credible and practical tool for making our factories more reliable, rather than a mere science-fiction concept.
📚 Sources :
- McCarthy et al., 1956 - A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence
- Krizhevsky et al., 2012 - ImageNet Classification with Deep Convolutional Neural Networks (AlexNet)
- Vaswani et al., 2017 - Attention Is All You Need (Transformers)
- McKinsey, 2024 - The state of AI: Global survey
- Statista, 2025 - Artificial Intelligence Worldwide Market Forecast