Where did AI come from?

1950s: the dawn of AI

AI as a formal discipline began in the 1950s, although its conceptual roots can be traced back even earlier. These conceptual roots could go back to the very first digital computers of the 1940s, to the publication of Frankenstein in 1818, or even all the way back to the Ancient Greek mythos of the ‘Bronze Robot’ Talos, defender of Crete. It is generally accepted that AI's true origins lay with British mathematician and logician Alan Turing who played a crucial role in laying the groundwork with his 1950 paper, “Computing Machinery and Intelligence,” which proposed the famous Turing Test to determine if a machine could exhibit human-like intellect.

The 1950s also saw the very first developments of AI programs, including the Logic Theorist, created by Allen Newell and Herbert A. Simon, which mimicked human problem-solving skills. This pioneering software was designed to mimic human problem-solving skills by proving mathematical theorems. It operated by applying a series of logical rules to transform the axioms and previously proven theorems into new theorems. The Logic Theorist successfully proved 38 of the first 52 theorems in Principia Mathematica, a foundational work in mathematical logic. Its creation marked a significant milestone in AI, demonstrating that machines could perform tasks requiring human-like reasoning.

1960s: early endeavours and optimism

The 1960s were marked by a surge of optimism and ambitious projects in AI. Researchers developed programs that could solve algebra problems, prove geometric theorems, and understand simple natural language. The introduction of the General Problem Solver by Newell and Simon exemplified the era's ambition to create universal problem-solving machines. Joseph Weizenbaum's ELIZA, an early natural language processing program, also emerged, simulating human conversation and showcasing the potential of AI in understanding and generating human language.

1970s: challenges and criticisms

The 1970s brought progress but also birthed a lot of scepticism and reality within the field. The limitations of early AI systems became apparent as they struggled with real-world complexities. The optimism of the 1960s had given way to a more critical and cautious approach. One major event that impacted AI research was the publication of the Lighthill Report in 1973. Commissioned by the British government and authored by Sir James Lighthill, the report criticised the lack of practical progress in AI and argued that the field had not lived up to its earlier promises. This led to a reduction in funding and interest in AI research, particularly in the United Kingdom, and initiated what became known as the "AI winter," a period of reduced funding and enthusiasm for AI research.

Despite these challenges, the decade saw advancements in expert systems, which were designed to mimic the decision-making abilities of human experts. Notable examples included MYCIN, a system developed at Stanford University for diagnosing bacterial infections and recommending antibiotics. MYCIN's ability to perform at a level comparable to human experts demonstrated the potential of AI in practical applications, even if the broader ambitions of the field remained unfulfilled.

Additionally, the 1970s saw progress in the field of robotics, with researchers exploring the integration of AI techniques into robotic systems. This period laid the foundation for future innovations in autonomous and intelligent robots, which would come to fruition in the following decades.

1980s: expert systems begin to rise

The 1980s marked a period of renewed enthusiasm and significant progress in artificial intelligence, largely driven by the development and commercialisation of expert systems. These systems, designed to emulate the decision-making abilities of human specialists, demonstrated practical applications of AI in various industries and led to increased investment and interest in the field.

Expert systems became the most successful AI technology of the decade. One prominent example was XCON (eXpert CONfigurer), developed by Digital Equipment Corporation (DEC) in collaboration with Carnegie Mellon University.

Additionally, the 1980s witnessed increased interdisciplinary collaboration, as AI researchers began to integrate insights from fields such as cognitive science, psychology, and neuroscience. This holistic approach enriched AI research and expanded its scope, leading to more sophisticated and human-like AI systems.

1990s: from logic to learning

The 1990s represented a significant shift in artificial intelligence from rule-based systems to learning-based approaches, driven by the increasing availability of data and advances in computational power. This decade saw a growing focus on machine learning, where algorithms were developed to enable systems to learn from and make predictions based on data.

One of the most notable milestones of the 1990s was the triumph of IBM's Deep Blue over world chess champion Garry Kasparov in 1997. This event highlighted the potential of AI in strategic decision-making and showcased the advancements in computational power and algorithm design. Deep Blue's victory was not just a technological achievement but also a moment of public recognition for the capabilities of AI.

The 1990s also saw significant advancements in natural language processing (NLP). Researchers developed more sophisticated techniques for understanding and generating human language, leading to improvements in applications such as speech recognition, translation, and information retrieval. These developments were driven by the availability of large corpora of text and speech data, as well as the increased computational resources to process this data.

In addition to these technical advancements, the 1990s were marked by the rise of the Internet, which played a crucial role in the proliferation of data. The expansion of the Internet facilitated the collection, sharing, and analysis of vast amounts of information, providing a rich resource for training machine learning models.

2000s: data-driven AI

The 2000s were a transformative period for AI, driven by the explosion of data and advances in computational power. The proliferation of the Internet and social media resulted in vast amounts of data, crucial for training sophisticated machine learning models.

Key developments included the rise of tech giants like Google, Amazon, and Facebook, which utilised AI to enhance search algorithms, recommendation systems, and content delivery. Natural language processing (NLP) saw significant progress, with techniques like Latent Dirichlet Allocation (LDA) improving language understanding and generation.

Support vector machines (SVMs) and kernel methods became effective tools for image recognition and bioinformatics. Robotics also advanced, with increased autonomy and applications in industrial automation and healthcare. The development of autonomous vehicles gained traction through DARPA Grand Challenge competitions.

Cloud computing emerged, providing the necessary computational resources for large-scale AI model training, and enabling broader access to AI technology. However, the decade also raised concerns about data privacy, security, and ethical considerations, highlighting the need for responsible AI deployment.

2010s: the era of deep learning

The 2010s were defined by the rapid advancement and widespread adoption of deep learning, which transformed artificial intelligence into a ubiquitous technology impacting various aspects of daily life. This era saw neural networks, particularly deep neural networks, achieving unprecedented success in tasks such as image and speech recognition, language translation, and autonomous driving.

One of the most significant milestones of the decade was the development of convolutional neural networks (CNNs), which excelled in image recognition tasks. A landmark achievement was Google's AlexNet winning the ImageNet competition in 2012, showcasing the power of deep learning. This success spurred a wave of research and development in AI, leading to further innovations and applications.

Another pivotal moment was the creation of AlphaGo by DeepMind, a subsidiary of Google. In 2016, AlphaGo defeated world champion Go player Lee Sedol, demonstrating the potential of AI in strategic thinking and complex problem-solving. This breakthrough underscored the capabilities of reinforcement learning, a type of machine learning where agents learn by interacting with their environment.

The decade also witnessed significant progress in natural language processing (NLP), driven by the introduction of transformer models. The release of models such as Google's BERT and OpenAI's GPT-2 and GPT-3 revolutionised NLP, enabling AI systems to understand and generate human language with remarkable accuracy. These advancements facilitated improvements in machine translation, sentiment analysis, and conversational agents.

2020s: AI in everyday life

The 2020s have seen artificial intelligence becoming deeply integrated into everyday life, with AI technologies continuing to evolve and expand their influence across various sectors. AI-powered devices and applications, such as virtual assistants, smart home systems, and autonomous vehicles, have become more common and sophisticated, improving efficiency and convenience.

One of the defining trends of the 2020s has been the advancement and widespread adoption of AI in healthcare. AI-driven technologies have been used for diagnostics, personalised medicine, and predictive analytics, enhancing patient care and medical research. For example, AI algorithms have been employed to detect diseases such as cancer from medical images with high accuracy, aiding early diagnosis and treatment.

The 2020s have also seen AI playing a critical role in addressing global challenges. AI has been leveraged for climate modelling, helping scientists better understand and predict climate change patterns. In agriculture, AI technologies have improved crop management and sustainability through precision farming techniques, which optimise resource use and increase yield.

The integration of AI with other emerging technologies, such as IoT and blockchain, has driven further innovation. AI-powered IoT devices have enabled smarter and more efficient cities, while blockchain has provided secure and transparent frameworks for AI applications in finance and supply chain management.

Ethical considerations have become increasingly prominent in the 2020s. Issues such as bias in AI algorithms, transparency, and accountability have been at the forefront of discussions among researchers, policymakers, and industry leaders. Efforts to develop explainable AI (XAI) have gained traction, aiming to make AI systems more transparent and understandable to users. Additionally, there has been a push for regulations and guidelines to ensure the responsible development and deployment of AI technologies.

In response to the growing influence of AI, education and workforce development have also been a focus. Initiatives to equip workers with the skills needed for an AI-driven economy have been implemented, emphasising the importance of continuous learning and adaptation in the face of technological advancements.

Looking ahead to what's next

Looking ahead, AI is poised to continue its rapid evolution. The integration of AI with other emerging technologies like quantum computing, the Internet of Things (IoT), and blockchain is expected to drive further innovation. AI's potential in areas such as personalised medicine, climate modelling, and advanced robotics promises to revolutionise various sectors. However, ethical considerations, regulatory frameworks, and the societal impact of AI will remain critical areas of focus to ensure that AI benefits all of humanity.

Commenting on the next step of AI, Georges-Olivier Reymond, Co-founder and CEO of Pasqal, says: "Last year marked a turning point where the world discovered the potential of generative AI (gen AI). Now, rapid adoption is driving real business benefits, like cost reduction and revenue growth. In fact, the global AI market is projected to reach $267 billion by 2027. But, what if AI could do even more? It’s time for AI to enter a new partnership with an emerging technology – quantum.

"By pairing these technologies into a hybrid model to create Quantum AI, it can enhance the efficiency and capabilities of AI algorithms and systems. At the moment, the primary bottleneck is the lack of computing power, while the very promise of quantum technology is increasing it at unprecedented levels. This translates into using quantum computers to accelerate AI training processes, optimise algorithms for specific tasks, or explore new approaches to machine learning and data processing that are enabled by quantum principles.

"Beyond this, Quantum AI tackles a crucial challenge: AI's growing energy footprint. As AI datasets and usage grows, so does its energy consumption. Pasqal’s flagship Orion quantum computer requires only 3kW, compared to 1,400+ kW for a classical supercomputer. With Quantum AI, energy efficiency becomes a reality allowing businesses to continue to harness AI’s potential, with a sustainable future in mind."

AI's journey from its conceptual origins to its current ubiquity in daily life has been marked by cycles of optimism, challenge, and breakthrough. As we continue to explore the possibilities of AI, it is essential to reflect on its history to understand its future trajectory and the profound impact it will have on society.