Introduction to AI - from the first steps to GenAI

Artificial Intelligence has fascinated us for more than two years, filling news headlines, social media and everyday conversations. While it may seem like a futuristic novelty, its story goes much deeper, creating a fascinating tale of human ambition to understand and recreate intelligence.

This is not my first approach to this topic. In July 2024, I wrote an article AI for Everyone or the Fascinating World of Artificial Intelligence  (in Polish) - my first blog post on AI for Everyone. Looking at it today, I feel a bit unsatisfied and recognize its beginner nature 🙂 Since then, I have read many new books, studied fascinating articles, and my knowledge has expanded significantly. Therefore, I decided to face this topic again.

This article is an invitation to travel together, but also my personal attempt to arrange in my head the knowledge I have gained. It is a journey through the history of AI, which will help us better understand the present and prepare for the future, and for myself to systematize the experiences and thoughts I have gathered. As someone who is constantly exploring the technology and sharing knowledge through the AI For Everyone project, I want to take you through the most important moments in the evolution of artificial intelligence, avoiding technical complexities and focusing on what really matters.

What exactly is artificial intelligence?

Defining AI is akin to trying to catch the wind in its sails - it is dynamic and changing. The European Commission describes AI as

"...systems that exhibit intelligent behavior, analyze their environment and take action - with some degree of autonomy - to achieve specific goals...."

Nils J. Nilsson of Stanford University puts it slightly differently:

"...artificial intelligence is the action of making machines intelligent, and intelligence is the characteristic that enables an individual to function properly and predictably in its environment...."

For the purposes of our journey, let's adopt a simpler definition:

AIs are systems that can learn, adapt and make decisions, mimicking human thought processes to some extent.

Fascinating journey through the history of AI

Beginnings and Foundations (1940s and 1950s)

Our journey begins in the fascinating period when humanity began its first attempts to understand and reproduce thought processes:

• First Neuron Models (1943)It all began with the groundbreaking work of Warren McCulloch and Walter Pitts, who presented the first models of artificial neurons. Their work not only inspired future researchers, but gave rise to the whole idea of creating neural networks that mimic the workings of the human brain.
  
  

McCulloch-Pitts neuron diagram. Source: Wikipedia

• Turing Test (1950) British mathematician Alan Turing proposed a thought experiment that still fascinates AI researchers today. The Turing test was designed to see if a machine could pretend to be a human in conversation convincingly enough to be indistinguishable from a real person. This test became a milestone in machine intelligence research.
• Dartmouth Conference (1956) In that landmark year, a conference was held at Dartmouth College that officially launched the era of artificial intelligence as a scientific discipline. It was there that the term "artificial intelligence" was used for the first time, and the meeting brought together prominent scientists from a variety of fields to work together on the development of AI.
• Pioneers of the discipline of AI (1956) John McCarthy, Marvin Minsky, Allen Newell and Herbert A. Simon - these are the names that will forever be written in the history of AI, establishing the foundations of the new scientific discipline.
• Birth of the Perceptron (1957) Frank Rosenblatt began his fascinating work on the Perceptron - one of the earliest neural network models. This model had an ambitious goal: to simulate the way human brains process information and learn from experience.
• Beginning of Machine Learning (1959) Arthur Samuel introduced the term "machine learning," presenting the revolutionary idea of programming computers that could learn to play checkers better than their creators. This was a breakthrough showing that machines can learn and improve their skills based on experience.

Era of First Breakthroughs (1960s and 1970s)

• 1966: ELIZA and DENDRAL The first systems capable of human communication (ELIZA) and expert decision support (DENDRAL) are created. Fascinatingly, you can still talk to ELIZA today by visiting https://web.njit.edu/~ronkowit/eliza.html - an amazing opportunity to experience one of the first chatbots ever in person!

Example of Eliza chatbot interface. 

• The first "AI winter" A period of disappointment and reduced funding shows that the road to true AI will be longer than initially expected.

Rebirth and New Directions (1980s)

The 1980s brought a fascinating renaissance in the field of AI, fueled by significant investment and development of new technologies:

• Development of Programming Languages Programming languages such as Prolog and Lisp gained popularity, which opened up new possibilities in creating increasingly complex AI systems. These tools became the foundation for subsequent breakthroughs in the field.
• Bayes networks (1985) Judith Pearl introduced the groundbreaking concept of Bayes networks - probabilistic structures representing causal relationships in the form of directed acyclic graph (DAG). This innovation significantly influenced the development of probabilistic inference systems.
• Back Propagation Algorithm (1986) This year brought a real revolution in the field of artificial neural networks. The application of the back propagation algorithm not only improved the performance of the network, but also opened the way for the development of deep machine learning. It was this algorithm that laid the foundation for the development of convolutional neural networks (CNNs), which today are indispensable in image, video and audio analysis.
• The Second "AI Winter" At the end of the decade, another "AI winter" occurred. Despite promising progress, disappointing results and funding cuts showed that we still faced enormous challenges. Researchers and investors faced not only conceptual limitations, but, most importantly, the technological barriers of the equipment at the time.

Renaissance and Breakthrough Achievements (1990s)

This decade brought a number of fascinating innovations and bold visions of the future:

• Visionary Perspective Vernor Vinge made a bold prediction in the 1990s - he predicted that within 30 years we would gain the technical means to create artificial intelligence superior to human intelligence. This vision, although it seemed futuristic, inspired many specialists to think more deeply about the future of AI.
• A.L.I.C.E. (1995) Richard Wallace created the chatbot A.L.I.C.E., which, using natural language and data from the Internet, opened up new possibilities for interactive human-machine communication.

Chatbot A.L.I.C.E. (Artificial Linguistic Internet Computer Entity)

• Breakthrough in Language Processing (1997) Sepp Hochreiter and Jürgen Schmidhuber developed the LSTM neural network architecture, which revolutionized the effectiveness of natural language processing models.
• Historic Moment in Chess (1997) IBM's Deep Blue computer defeated world champion Garry Kasparov - an event that became a symbol of AI's growing analytical capabilities and demonstrated that machines can compete with humans even in the most challenging fields.

The Era of Big Data and Deep Learning (2000-2010)

This period brought fundamental changes in the approach to AI:

• Big Data Revolution The years 2000-2010 were a time when huge datasets available on the Internet began to drive the development of deep neural networks, opening up entirely new possibilities.
• Birth of the GAN (2004) Fascinating discussions on generative inter-organizational networks (GANs) began, ushering in an era of AI-powered creativity.

Show how GANs work (photo: TNW)

• Breakthrough in Deep Learning (2006) Geoffrey Hinton revolutionized the approach to machine learning, giving rise to the era of deep learning. His work on back propagation opened a whole new chapter in the history of AI.

Dynamic Progress (2010-2020)

This decade has brought a series of breakthroughs:

• Watson and Jeopardy! (2011) IBM's Watson supercomputer won the "Jeopardy!" quiz show, demonstrating impressive advances in natural language processing.

Ken Jennings and Brad Rutter, the leading champions of the quiz show "Jeopardy!", faced off against a computer named Watson.
Source photo: Carol Kaelson/Jeopardy Productions Inc. provided by the Associated Press.

• The CNN Revolution (2012) The development of convolutional neural networks has revolutionized image analysis, paving the way for new AI applications.
• Era GAN (2014) Ian Goodfellow and his team defined the concept of GANs, introducing a revolutionary tool in the field of machine learning.
• Strategic Game Breakthrough AlphaGo (Google) in 2014 perfected reinforcement learning models, and in 2016 defeated Lee Sedol, one of the best go players, by a score of 4:1 - a moment that went down in AI history.
• New Horizons.
  • Birth of OpenAI (2015) with a mission to develop AI for the benefit of humanity
  • WaveNet (2016) generating speech nearly indistinguishable from human speech.
  • Implementation of transformer architecture by Google (2017)
  • First pretreated GPT-1 generative model (2018)

Era of Generative Artificial Intelligence (After 2020)

• GPT-3 (2020) OpenAI has introduced the GPT-3 model, which has revolutionized the way machines communicate with people.
• DALL-E (2021) A tool for generating images from text descriptions was created, opening up new possibilities in the field of creative AI.
• ChatGPT Breakthrough (2022) The release of ChatGPT became a milestone in the history of AI, revolutionizing human-machine interaction. This year went down as a breakthrough in the context of generative AI.

Logo of ChatGPT and OpenAI

• New Generation (2023-2024) More innovations have been introduced: ChatGPT-4o and o1, Gemini (formerly BARD) from Google, Bing AI from Microsoft and Claude AI from Anthropic. So many AI solutions were developed between 2023 and 2024 that they deserve a separate, detailed article.

Where are we going?

The history of AI is a story of constant development, full of ups and downs, but always moving toward a future in which technology is becoming an increasingly integral part of our lives. Each stage of this journey has brought new discoveries and possibilities, changing our understanding of what is possible.

As witnesses and participants in this technological revolution, we have the privilege of not only observing these changes, but also actively participating in them. Artificial intelligence is no longer the domain of scientists and programmers - it has become a tool available to anyone who wants to learn about its capabilities.

What will the future bring?

• Further development of generative models
• Increased availability and democratization of AI technologies
• New applications in medicine, education and other fields
• Increasing integration of AI into everyday life

Call for further exploration

This journey through the history of AI is just the beginning. I encourage you to continue to explore the topic, experiment with the tools available and share your experiences. Remember - AI is for everyone, and our shared adventure with this fascinating technology is just beginning.

This article is part of the AI For Everyone project, whose mission is to democratize knowledge about artificial intelligence. I invite you to keep an eye out for future publications and share your thoughts in the comments.

I invite you to follow us in the comments.

Note!
This article was prepared with the support of Claude 3.5 Sonnet, an advanced AI language model that helped organize and present content, while maintaining authentic message and educational value.

This article has also been automatically translated from Polish using DeepL. If you find any errors, please let me know in the comments.