History of Artificial intelligence in Timeline

By Popular Timelines Editorial Team · Updated: July 12, 2026

Artificial Intelligence (AI) refers to the simulation of human intelligence processes by computer systems. These processes include learning, which involves the acquisition of information and rules for using it; reasoning, the use of rules to reach approximate or definite conclusions; and self-correction. AI technologies leverage vast datasets and complex algorithms—most notably machine learning and deep learning—to identify patterns, make predictions, and automate decision-making across diverse fields. From natural language processing and computer vision to autonomous vehicles and generative models, AI is transforming industries by enhancing productivity, efficiency, and problem-solving capabilities. However, its rapid advancement also necessitates careful consideration of ethical frameworks, data privacy, and the socio-economic impacts of automation. As a transformative technology, AI continues to evolve, pushing the boundaries of what machines can achieve and fundamentally altering how humans interact with the digital and physical world.

1956: Founding of Artificial Intelligence as an Academic Discipline

In 1956, artificial intelligence was officially established as an academic field of study, marking the beginning of its formal history.

1979: Three Mile Island Nuclear Meltdown

In 1979, the Three Mile Island nuclear facility experienced a significant safety failure, specifically a partial nuclear meltdown of its Unit 2 reactor.

May 1997: Deep Blue Defeats Garry Kasparov

On May 11, 1997, the IBM computer system Deep Blue made history by becoming the first machine to defeat a reigning world chess champion, Garry Kasparov, in a competitive match.

2011: Watson Wins Jeopardy!

In 2011, IBM's question-answering AI system, Watson, demonstrated its capabilities by decisively defeating legendary Jeopardy! champions Brad Rutter and Ken Jennings during a televised exhibition match.

Loading Video...

2012: The Rise of Deep Learning

Beginning in 2012, deep learning experienced a sudden surge in success driven by increased computational power from GPUs and the utilization of massive datasets like ImageNet.

2012: Surge in AI Research via Deep Learning and GPU Acceleration

Starting in 2012, interest and funding in artificial intelligence grew substantially as researchers began utilizing graphics processing units to accelerate neural networks, enabling deep learning to surpass previous AI performance benchmarks.

2015: Peak Deep Learning Expansion

By 2015, the period of rapid success for deep learning continued, solidified by the widespread availability of high-speed GPU computing and large-scale training data.

March 2016: AlphaGo Defeats Lee Sedol

During March 2016, the AlphaGo AI system achieved a major milestone by winning 4 out of 5 games against professional Go champion Lee Sedol, marking the first time a computer defeated a pro player without a handicap.

2016: Shift in Privacy Perspective

In 2016, privacy experts like Cynthia Dwork began re-evaluating privacy through the lens of fairness, while Brian Christian noted a shift in focus from merely asking what data is collected to questioning how that data is being utilized.

2017: AlphaGo Defeats Ke Jie

Building on its previous success, the AlphaGo AI system defeated Ke Jie, the top-ranked Go player in the world, in a 2017 match.

2017: Widespread AI Adoption in Corporate Processes

During a 2017 survey, data indicated that approximately 20 percent of businesses had successfully integrated Artificial Intelligence into their existing products or operational processes.

2017: Advancement Through Transformer Architecture

The growth and development of artificial intelligence technologies experienced further acceleration beginning in 2017 with the introduction and implementation of the transformer architecture.

2019: AlphaStar Reaches Grandmaster Level

In 2019, DeepMind’s AlphaStar AI system demonstrated high-level performance by reaching the grandmaster rank in the real-time strategy game StarCraft II.

2019: Emergence of Generative Pre-trained Transformer Models

In 2019, generative pre-trained transformer (GPT) language models marked a significant advancement in artificial intelligence by demonstrating the ability to generate coherent and fluid natural language text.

2021: Unexplained Deep Learning Success

As of 2021, the underlying reasons for the exceptional performance of deep learning models across numerous artificial intelligence applications remain scientifically unknown.

2021: AI Agent Wins Gran Turismo Competition

In 2021, an AI agent utilizing deep reinforcement learning successfully competed against and defeated four of the world's top human players in a Gran Turismo driving competition.

2021: AlphaFold 2 Protein Structure Breakthrough

In 2021, the AlphaFold 2 system demonstrated a significant advancement in computational biology by successfully approximating the 3D structure of proteins in a matter of hours, a task that previously required months of laboratory work.

2022: Historical US Data Center Power Consumption

In 2022, US data centers accounted for 3% of the total US power consumption, serving as the baseline for the significant growth projected by the end of the decade.

2022: Closure of Palisades and Passage of Inflation Reduction Act

In 2022, the Palisades Nuclear reactor was closed, while simultaneously, the US government passed the Inflation Reduction Act, which provided critical tax incentives for nuclear energy projects.

February 2023: Introduction of Bing Chat

In February 2023, Microsoft officially launched Bing Chat, a service later rebranded as Copilot Search, which utilizes artificial intelligence to generate concise summaries for user queries.

September 2023: Final Approval Phase for Data Centers

In September 2023, Taiwan reached a milestone by issuing the final approvals for data center projects before shifting its policy stance regarding regional infrastructure.

2023: AI Performance on Standardized Testing

By 2023, advanced transformer-based AI models achieved human-level performance scores on various high-stakes examinations, including the bar exam, the SAT, and the GRE, showcasing their widespread utility in real-world applications.

2023: AI-Guided Antibiotic Discovery

During 2023, it was reported that artificial intelligence was utilized to facilitate drug discovery, resulting in the identification of a new class of antibiotics capable of neutralizing two distinct strains of drug-resistant bacteria.

2023: Authors Sue AI Companies for Copyright Infringement

In 2023, prominent authors including John Grisham and Jonathan Franzen initiated legal action against various artificial intelligence companies, alleging that these firms utilized their copyrighted literary works to train generative AI models without authorization.

January 2024: IEA Electricity 2024 Report Release

In January 2024, the International Energy Agency published a report titled Electricity 2024, Analysis and Forecast to 2026, which marked the agency's first attempt to project the impact of AI and cryptocurrency on global power consumption.

March 2024: Amazon nuclear-powered data center acquisition

In March 2024, Amazon finalized the purchase of a nuclear-powered data center located in Pennsylvania for a total of US$650 million as part of a larger trend of AI firms securing reliable energy sources.

September 2024: Microsoft Agreement to Reopen Three Mile Island

In September 2024, Microsoft entered into a 20-year agreement with Constellation Energy to reopen the Three Mile Island nuclear power plant, aiming to secure 100% of the energy produced to power its operations.

2024: AIME 2024 Mathematical Benchmark

During 2024, the American Invitational Mathematics Examination (AIME) served as a benchmark for evaluating the mathematical reasoning capabilities of artificial intelligence models, such as those enhanced by the rStar-Math technique.

2024: Suspension of Data Center Approvals in Taoyuan

During 2024, the Taiwanese government implemented a suspension on the approval of data centers north of Taoyuan that exceed a 5 MW capacity, primarily due to concerns regarding local power supply shortages.

2024: Google DeepMind Launches SIMA

In 2024, Google DeepMind unveiled SIMA, an AI agent designed to play open-world video games autonomously by processing visual screen data and following natural language commands.

2024: Goldman Sachs Report on AI Data Center Power Demand

In 2024, a Goldman Sachs Research Paper titled AI Data Centers and the Coming US Power Demand Surge was published, highlighting that artificial intelligence infrastructure is driving an unprecedented surge in US electrical power demand.

2024: Emergence of Reasoning Models

In 2024, a specialized category of large language models known as reasoning models gained prominence. These models are uniquely trained to produce intermediate chains-of-thought, which has led to significant advancements in solving complex mathematics and coding problems.

2024: Machine Learning in Parkinson's Disease Treatment

In 2024, researchers applied machine learning to expedite the search for Parkinson's disease treatments by targeting alpha-synuclein protein aggregation, achieving a ten-fold increase in screening speed and a thousand-fold reduction in costs.

2024: AI Usage in 2024 Indian Elections

Throughout the 2024 Indian elections, political entities invested US$50 million into AI-generated content. This initiative included the creation of deepfakes featuring current and deceased politicians to improve voter engagement, as well as the use of AI to translate political speeches into multiple regional languages.

January 2025: Microsoft Proposes rStar-Math Technique

In January 2025, Microsoft introduced rStar-Math, an innovative technique that integrates Monte Carlo tree search with step-by-step reasoning. This approach allows a smaller model like Qwen-7B to achieve high performance, specifically solving 53% of the AIME 2024 problems and 90% of the MATH benchmark problems.

April 2025: Creation of the AI Hallucination Cases database

In April 2025, Damien Charlotin, a lecturer at HEC Paris and Sciences Po, established the global AI Hallucination Cases database to track instances where generative AI tools provided inaccurate information for legal research.

May 2025: Google Introduces AI Mode

During the Google I/O event held in May 2025, Google officially unveiled a new feature referred to as AI Mode, marking a significant advancement in their product lineup.

October 2025: Planned Reopening of Palisades Nuclear Reactor

As of October 2025, the Palisades Nuclear reactor on Lake Michigan is slated to be reopened following significant financial investment from the US government and the state of Michigan.

2025: Nuclear Power Phase-out Goal

As of the targeted year 2025, Taiwan is working toward the national objective of phasing out nuclear power generation.

2025: McKinsey & Company AI Infrastructure Report

In 2025, the consulting firm McKinsey & Company released a report projecting that the United States would see a massive $2.7 trillion investment into artificial intelligence infrastructure and data centers by the year 2030, a spending rate that would exceed the monthly costs of the World War II Manhattan Project.

2025: Gemini Deep Think achieves gold medal results at the International Math Olympiad

In 2025, the experimental model Gemini Deep Think demonstrated significant progress in mathematical reasoning by utilizing natural language prompts to solve complex tasks, ultimately earning gold medal results at the International Math Olympiad.

2026: Projected Surge in Energy Demand

By 2026, the International Energy Agency projects that the electricity required to power artificial intelligence and cryptocurrency operations could double, potentially reaching a level of total consumption equivalent to that of Japan.

2026: Judicial sanctions and bar association warnings regarding AI hallucinations

By 2026, the legal system faced significant challenges as judges began issuing formal sanctions and bar associations released warnings after attorneys submitted court filings containing fabricated case law generated by AI tools.

2026: Dozens of AI disease-prediction models were trained on dubious data

In 2026, a Nature article revealed that 125 research papers utilized unreliable data to train artificial intelligence models for predicting conditions such as diabetes and stroke. The report raised concerns as some of these AI-based medical tools were already implemented in patient care, though it remained unconfirmed whether these instances resulted in negative health outcomes.

2030: Projected US Data Center Power Consumption

By 2030, it is forecasted that US data centers will account for 8% of the total US power consumption, creating significant pressure on the electrical grid and potentially necessitating expansion in power generation.

2030: Projected AI Infrastructure Investment Milestone

By the year 2030, the United States is expected to reach a cumulative investment of $2.7 trillion in AI-related infrastructure and data centers, representing a historic shift in capital allocation toward artificial intelligence technologies as forecasted in 2025.