Artificial Intelligence: History, Evolution, and Impact
A factual archive, explored as a knowledge graph rather than a wall of text.
Artificial intelligence is the effort to build machines that do
the things we call intelligent when a person does them — reasoning, learning,
perceiving, and using language. The phrase was coined in 1955, and the field was
organized a year later at the Dartmouth conference in 1956. Everything below is a
way into that story: hover, click, drag, and expand your way through it.
Knowledge graph
One central idea, three branches. Start here.
Hover a branch to highlight it · click a branch head (+) to expand
AI timeline
Seventy years, unfolding as you scroll.
1950
Turing's “Computing Machinery and Intelligence”
The question posed: can machines think?
Idea The imitation game — the “Turing test”
Author Alan Turing
Legacy A concrete test for machine intelligence
1956
Dartmouth Conference
Artificial intelligence is named as a field
Organizer John McCarthy
With Minsky, Rochester, Shannon
Claim Intelligence can be precisely described
Legacy AI becomes a discipline
1965
DENDRAL — an early expert system
Encoding a specialist’s rules in software
Where Stanford
Idea Rules capture expert chemistry knowledge
Legacy The 1980s expert-systems boom
1974
The first AI winter
Funding retreats as promises go unmet
Cause Symbolic systems were brittle, did not scale
Pattern Hype, then retreat — it recurs
1986
Backpropagation popularized
A practical way to train multi-layer networks
Figures Rumelhart, Hinton, Williams
Idea Learn features by propagating error
Legacy Foundation of modern deep learning
1997
Deep Blue beats Kasparov
A machine defeats the world chess champion
Builder IBM
Method Massive search plus evaluation
Legacy Focused machines can beat human experts
2012
AlexNet
Deep learning breaks image recognition
Figures Krizhevsky, Sutskever, Hinton
Fuel GPUs + the ImageNet dataset
Legacy The deep-learning revolution
2017
“Attention Is All You Need”
The transformer architecture arrives
Lab Google
Idea Attention replaces recurrence
Legacy The architecture behind modern LLMs
2022
Generative AI goes mainstream
Large language models reach the public
Shift Chat interfaces for general use
Reach Hundreds of millions of users
Legacy AI becomes everyday software
2026
The present
Where the record stands today
State Capable at tasks — not thinking
Debate Genuine benefits, genuine risks
Open Does understanding ever follow?
Scroll down to move through the years
AI family tree
How the approaches branch — symbolic, statistical, deep.
▶Artificial Intelligence
▶Symbolic AI
▶Logic
▶Planning
▶Expert Systems
▶Machine Learning
▶Regression
▶Decision Trees
▶SVM
▶Neural Networks
▶Deep Learning
▶CNN
▶RNN
▶LSTM
▶Transformer
Click a branch to expand or collapse it
AI evolution graph
The lineage that leads to the transformer. Drag the nodes; click to explain.
Obsidian-style graph
Drag the nodes
Rearrange the lineage, then click any node to read what it is and when it arrived.
Relationship graph
People and the field, and how they connect.
Click a node to see how it connects.
AI museum
A walk through the landmark years.
Hall 1
1943
The McCulloch–Pitts neuron
The first mathematical model of how a neuron might compute — logic from biology.
Warren McCulloch · Walter Pitts
Hall 2
1950
Turing’s imitation game
Turing reframes “can machines think?” as a test anyone can run.
Alan Turing
Hall 3
1956
The Dartmouth conference
Artificial intelligence is named and organized as a field of research.
McCarthy · Minsky · Rochester · Shannon
Hall 4
1980
The expert-systems era
Rule-based AI reaches industry, encoding specialists as if–then rules.
Feigenbaum and the Stanford school
Hall 5
2012
AlexNet
Deep learning breaks image recognition, powered by GPUs and ImageNet.
Krizhevsky · Sutskever · Hinton
Hall 6
2022
Generative AI goes public
Large language models move from the lab into everyday conversation.
Many labs
Scroll sideways to walk through the halls →
Read the archive
The full essays behind the graph.
John McCarthy John McCarthy (1927–2011): the scientist who coined 'artificial intelligence,' organized the 1956 Dartmouth conference, and created LISP.
Evolution & Timeline A decade-by-decade history of AI: 1950s foundations, symbolic AI, expert systems, machine learning, deep learning, and modern LLMs.
Benefits Where AI has produced measurable value: automation, medical imaging, large-scale data analysis, adaptive education, and scientific research.
Harms and Risks Documented harms of AI: job displacement, dataset bias, lack of interpretability, surveillance misuse, deepfakes, and overdependence.
Philosophy The philosophy of AI: symbolic versus data-driven intelligence, the limits of today's systems, and whether machines can truly understand.
Conclusion What the record shows: AI has evolved over decades through incremental progress, and its genuine benefits and genuine risks must be weighed together.
AI Topics glossary
Plain-language explainers for 100+ of the most-searched AI terms —
ChatGPT, LLMs, transformers, AGI, and more.
Blog
Essays and notes on the history, ideas, and impact of AI.
About this archive
The purpose of the archive and the authentic sources it draws on.