When someone asks who invented AI, it sounds at first like a typical quiz question on *Günther Jauch* with exactly one correct answer. Just like “Who invented the telephone?”—name, year, done.
With AI, it’s different: AI isn’t a single device, but an idea, a field of research, and a long chain of breakthroughs. And yes: there are a few names in that chain without whom we wouldn’t have chatbots, image recognition, or smart assistant systems. We’ll trace the development for you.
The most important information in brief
- Who invented AI? No one person alone. AI is the result of many minds, ideas, and technologies over the course of decades.
- The term “artificial intelligence” (AI) is generally attributed to John McCarthy—through the 1956 Dartmouth Workshop (or the 1955–56 proposal).
- In 1950, Alan Turing introduced one of the central concepts with the “Imitation Game” (Turing Test): making intelligence testable through behavior.
Who invented AI? An overview of the most important developments
Before we dive into dates and names, here’s a bit of a reality check: The question “Who invented AI?” isn’t the story of a single hero, but rather a relay race spanning decades.
At times, AI was conceived as a set of rules (“If X, then Y”), at other times as a learning system that teaches itself patterns, and in between there were periods when the industry was essentially in hibernation.
That’s exactly why it’s worth taking a look at the key milestones: You’ll see how AI development has progressed from fundamental philosophical ideas through mathematical models to today’s breakthroughs. And suddenly, you’ll be able to answer the question of when AI first emerged in your sleep.
1956: Dartmouth – the birth of “Artificial Intelligence”
The “Dartmouth Summer Research Project on Artificial Intelligence” is widely regarded as the founding event of AI as a field of research. It was organized by John McCarthy, Marvin Minsky, Claude Shannon, and Nathaniel Rochester, among others. Dartmouth itself notes that the term “Artificial Intelligence” was coined there and that the conference is considered the “birth” of the field.
So if you had to answer in a single sentence who “invented” AI, John McCarthy is the name that comes up most often, because he coined the term and defined the framework. But a term alone does not make for a functioning AI. It took more than that.
1950: Alan Turing – the idea of making intelligence testable
Even before Dartmouth, Alan Turing wrote in his 1950 paper “Computing Machinery and Intelligence” posed the question “Can machines think?” and proposed the “Imitation Game” (later known as the “Turing Test”). This was significant because Turing thereby shifted the focus: away from semantic quibbling (“What is thinking?”) toward observable behavior.
1957–1960: Early neural networks – the perceptron as a precursor
While symbolists were thinking about rules, early attempts to replicate learning were taking place in parallel. Starting in 1957, Frank Rosenblatt worked on the Perceptron, an early (very simple) learning model. This wasn’t yet “modern AI,” but it shows that AI development followed a dual-track approach from the very beginning, with rules and learning proceeding simultaneously.
The 1960s to the 1980s: Symbolic AI, expert systems—and disillusionment
In the 1960s and 1970s, the idea of modeling intelligence using symbols, logic, and rules was all the rage. Later came expert systems: knowledge was modeled manually. This worked in narrow domains until it became too expensive, too fragile, and too difficult to scale. Then, as so often happens in tech history, a long winter set in.
AI Winter: When Progress Freezes
“AI Winter” refers to periods during which interest in and funding for AI declined significantly, particularly following overly ambitious promises and insufficient actual performance. This serves as a good reminder today: AI is powerful, but it remains a system with limitations, risks, and dependencies (data, quality, governance).
1986: Backpropagation – the training algorithm becomes practical for everyday use
A major step forward for machine learning systems was the widespread adoption of backpropagation (propagating errors backward through the network to adjust weights). The classic Nature article by Rumelhart, Hinton, and Williams (1986) is one of the key references.
That is one of the reasons why names like Geoffrey Hinton come up so often these days in discussions about “who invented AI”: he played a major role in shaping the modern era of machine learning and neural networks.
2012: AlexNet – Deep learning shows its true potential
In 2012, AlexNet (Krizhevsky, Sutskever, Hinton) won the ImageNet competition by a wide margin—a wake-up call that deep networks combined with GPU power could suddenly scale. From that point on, AI was no longer just “research” in many companies, but a product.
2017: Transformers – The Foundation of Today's Language Models
The next major breakthrough came in 2017 with “Attention Is All You Need”: the Transformer.
Transformer models are (in simple terms) extremely good at learning relationships within sequences and form the backbone of many modern language and multimodal models. When people ask today who invented AI, what many actually mean is: Who made modern generative AI possible? And 2017 is a pivotal year in that regard.
From that point on, AI didn’t just get “better”—it suddenly became something the public could experience firsthand: models like BERT (2018) demonstrated just how powerful transformers could become in language tasks, GPT-3 (2020) made large-scale text generation widely accessible, and with ChatGPT (2022), the technology finally became a part of everyday life.
The Future of AI – From Tool to Team Member
The coming years will be less about “bigger and smarter” and more about integration: AI as part of processes, roles, and chains of responsibility.
In practical terms, this means:
- Data quality is becoming the bottleneck (not the choice of model).
- Security and compliance are moving from the "to-do later" list to the very top.
- Use cases must be measurable: time, quality, risk, revenue—something concrete.
AI is becoming more specialized (and therefore more useful)
Instead of one-size-fits-all models, we are seeing more domain-specific systems and solutions: AI for support, AI for quality control, AI for knowledge management, AI for documents. When people ask, “Since when has AI existed?” they often mean, in a business context: Since when has AI become truly relevant to us? For many, the answer is: since it became possible to integrate it into everyday life and since it became economically viable.
AI is certainly capable of a lot. But its true value only becomes apparent when someone effectively bridges the gap between IT, business units, and strategy. That is precisely where the line is drawn between a demo and real value creation.
AI is a development, not an invention
“Who invented AI?” is the wrong question, but it stems from the right curiosity. AI isn’t the work of a single inventor; it’s a relay race: Turing provided the conceptual framework, Dartmouth gave the field its name, McCarthy coined the term, Rosenblatt and colleagues developed early learning models, backpropagation made training more practical, AlexNet brought about the practical breakthrough, and Transformers ignited the current generation.
When you talk about this within the company, the bottom line is this: AI has grown—and it continues to grow. The difference isn’t down to magic, but to clear goals, good data, and a plan that takes implementation and accountability into account.
This is exactly where BE BRAVE comes in: as a partner that not only explains AI but also integrates it into processes, teams, and systems in a way that turns it into a real business lever—pragmatic, well-structured, and with an eye toward Swiss standards, data protection, and feasibility.
Frequently Asked Questions
Who invented AI—John McCarthy or Alan Turing?
If you had to name one person: John McCarthy coined the term “artificial intelligence” and organized the 1956 Dartmouth Workshop. In 1950, Alan Turing laid important groundwork for the concept of machine intelligence and the Turing test.
How long has AI been around?
As a field of research, it is generally considered to have begun “officially” in 1956 (Dartmouth). As an idea (machines that “think”), it goes back much further—but 1950 and 1956 are the standard reference points.
Why was there an AI winter?
Because expectations and reality sometimes diverged: overly ambitious promises, insufficient computing power, insufficient data, and fragile systems. The result: less funding and interest.
What was the most significant breakthrough in modern AI?
There were several, but two milestones are essential: AlexNet in 2012 (which popularized deep learning) and Transformer in 2017 (which forms the basis for many of today’s language models).