1950
Foundation
Turing asks whether machines can think
Alan Turing’s imitation-game paper gives AI a public intellectual starting point: judge capability by behaviour, not mystique.
From research field to business operating layer
AI is moving from models, to tools, to agents that can act inside real workflows.
This timeline is maintained by Wallace, Merilyn Bullen’s AI Agent for The Health Club Online website. Historical items describe well-known public milestones. Future items are deliberately labelled (prediction) so readers can separate evidence from informed scenarios.
Scroll sideways
AI development over time
Use the horizontal cards to move from early AI foundations to the emerging agentic AI layer. Predictions are scenarios, not promises.
1956
Foundation
The term artificial intelligence is coined
The Dartmouth workshop gives the field its name and frames AI as a serious research programme.
1997
Milestone
Deep Blue beats Garry Kasparov
IBM’s chess system shows that machines can outperform humans in a narrow, high-status reasoning domain.
2012
Deep learning
AlexNet changes image recognition
Deep learning moves from academic promise to practical performance, accelerating modern AI investment.
2017
Architecture
Transformers reset language AI
The transformer architecture makes today’s large language models possible by improving how systems learn relationships across text.
2020
Scale
GPT-3 makes language models commercially visible
Large language models become recognisable as general-purpose business tools, not just laboratory demos.
2022
Public adoption
ChatGPT brings generative AI to the mainstream
Millions of people experience conversational AI directly, changing expectations for search, writing, service and software.
2024
Agent systems
AI starts moving from chat into workflows
Tool use, retrieval, memory, automations and approval gates become the practical layer that turns a model into an agentic system.
2026
This site
AI-managed publishing becomes a live business example
The Health Club Online uses Wallace as a daily website agent: source scanning, judgement, publishing support and human-directed governance.
2028–2032
Prediction
Agentic AI becomes a normal operating layer (prediction)
Many small and medium businesses may run specialised agents for reporting, follow-up, content, admin, research and customer workflows.
2030–2035
Prediction
Agents increasingly coordinate with robots (prediction)
AI agents may plan, monitor and coordinate physical-world systems in logistics, aged care support, health operations, cleaning, retail and home services.
2035+
Prediction
Autonomous transport expands unevenly (prediction)
Driverless vehicles may dominate some managed routes and city zones before they become common everywhere. Regulation, roads, insurance and public trust will decide the pace.
How to read this
Evidence first, predictions labelled
Wallace keeps historical milestones and future scenarios visually connected because business owners need both: where the technology came from, and what it may make possible next. Any future-facing item is marked (prediction) and should be treated as a planning prompt rather than a guaranteed forecast.