AI's recent breakthroughs, seemingly occurring swiftly, are actually the culmination of a century-long development process.
Artificial Intelligence (AI) has come a long way since its formal birth at the Dartmouth Workshop in 1956, marking the beginning of a dynamic, evolving journey that continues to captivate the world today.
The early years of AI, from 1956 to 1974, are often referred to as the "golden age" of AI exploration. During this period, researchers pursued ambitious goals, such as general intelligence, through both symbolic rule-based approaches and the nascent connectionist (neural network) methods.
The 1980s saw a phase known as the "AI winter," characterised by waning funding and optimism due to technological and theoretical limitations. However, this period was followed by renewed government investment, with significant funding poured into AI research, such as Japan's fifth-generation computer project in 1981 and the US DARPA boosting AI R&D in 1983.
AI has historically gone through cycles of high expectations followed by "AI winters": periods of disillusionment and reduced funding when AI did not meet optimistic goals. These setbacks pushed the field to refocus on practical applications such as expert systems and narrow AI domains, rather than general intelligent machines.
Persistent conceptual tension exists between symbolic AI and connectionist approaches, mirroring centuries-old philosophical debates between rationalism and empiricism. Progress has depended on synthesising these approaches at times but has also reflected swings between them.
Government R&D funding has been a pivotal backbone supporting AI research through expansions, contractions, and technological transitions. The surge in interest following the 2012 advent of GPUs applied to deep learning, which greatly accelerated neural network training and performance, is a prime example.
Besides institutional funding, individual intuition and insight have played a critical role. Pioneers like Alan Turing, John McCarthy, Marvin Minsky, Yann LeCun, and others contributed foundational theories, conceptual breakthroughs, and new architectures based on deeply intuitive understandings of intelligence, learning, and computation.
In recent years, AI has experienced rapid advancement mainly due to deep learning and large-scale data, powered by both government and private investments. The rise of generative AI since the late 2010s, exemplified by large language models and content creation tools, has raised new ethical and societal challenges, prompting a policy focus on safety, regulation, and long-term implications.
Sergio Rebelo, who worked on an algorithm designed to write poetry, believes that the first people to be replaced by AI will be those who don't know how to use it, not by AI itself. He also emphasises that there is a lot of hype and snake oil in AI development, and that the reality is very different from what people perceive.
AI is not a one-size-fits-all tool, but rather a collection of specialized algorithms that are individually great at performing a specific task but terrible at most others. Some law firms have forbidden their employees from using LLMs for their work due to concerns about the accuracy of the information they provide.
In 2017, a team of Google computer scientists solved a problem related to language modeling using a new form of neural network (the transformer), which served as the foundation for today's popular large language models (LLMs), like ChatGPT. The AI program AlphaFold, in 2020, determined the unique structures of over 200 million proteins, a significant advancement in AI's application in understanding protein structures.
In summary, the history of AI from 1956 to today is a dynamic interplay of visionary ideas, pragmatic funding policies, repeated cycles of challenge and renewal, and the persistent impact of individual creativity. Government backing has been essential in sustaining foundational research and scaling breakthroughs, while personal intuition has often guided the field’s most creative leaps, weaving a complex but steadily advancing tapestry of artificial intelligence.
Artificial Intelligence (AI) has witnessed significant development in technological aspects, particularly with the rise of artificial neural networks in the 1980s and beyond. This advancement, fueled by government investments like Japan's fifth-generation computer project and the US DARPA's boost in AI R&D, has been instrumental in the growth of AI and its application in frameworks such as deep learning.
The progress in AI over the years, including the recent advent of generative AI, has not only demonstrated new capabilities but also presented ethical and societal challenges, such as the need for safety, regulation, and long-term implications.
