BRAVING THE NEW WORLD OF AI
Amin Saifan (Year 11) considers the history of Artificial Intelligence, and wonders what the future holds
F or decades, researchers have sought to create intelligent machines that can replicate facets of human cognition, an endeavour filled with bold visions, frustrating setbacks and patient perseverance. The past three years have seen us entering a new era, with artificial intelligence going through a period of explosive growth. We are witnessing the genesis of what is being referred to as the AI Age, a monumental turning point in human history, with artificial intelligence set to transform civilization in the decades to come. In February, NVIDIA unveiled an AI agent named MAX that can perform real-time simultaneous translation between multiple languages, while accurately mimicking the voices of historical figures The groundwork for this recent period of rapid progress was laid down by pioneering developments over the past few decades. Key waypoints included the inception neural network (this was the AI which included Deep Blue, later known as Alpha Zero, and which was able to beat Gary Kasparov) and leaps forward in machine- learning during the 1980s and 1990s, with algorithms being developed which heralded the arrival of home computing and gaming. The emergence of big data and modern parallel computing in the 2000s enabled further major steps forward in deep-learning algorithms, giving us, for example, reactive search engines such as Google. However, following many triumphs at the end of the 20th century, the early 2000s was a period of relative stagnation, with AlphaZero’s retirement to medicine development, and the horror of Microsoft’s AI Twitter
bot that highlighted the dangers of AI: after 48 hours of exposure to the human race it attempted to incite a race war, and was withdrawn. It was only when ChatGPT’s capabilities were made available to the public that we witnessed the ‘comeback’ of AI, with systems like DeepMind’s Gato and Anthropic’s Claude exhibiting startling aptitude at multitask reasoning and natural language processing. Google and Anthropic stunned the AI community by unveiling systems that could generate coherent, thoughtful prose, rivalling human creativity. OtherslikeYou.com showed the ability to synthesize complex concepts just from natural language prompts. Last year, we witnessed a barrage of further break- throughs thanks to Moore’s Law (which allows for expo- nential growth) fuelling soaring model size and voracious data consumption. Systems like DeepMind’s Sparrow were now ingesting billions of webpages to produce essays, to solve maths problems, and to generate code. Startups began deploying AI bots which could pass stand- ardized tests and summarize complex documents across scientific and legal domains. Thus far, at time of writing halfway through 2024, the pace of progress shows no signs of slowing. In February, NVIDIA unveiled an AI agent named MAX that can per- form real-time simultaneous translation between multiple languages, while accurately mimicking the voices of his- torical figures. More demos emerge daily of AIs that can generate award-winning music, concept art and product designs, just from short text prompts. Artificial Intelligence is already transforming how we work, create and communicate. Nearly every industry, from healthcare to transportation, is rushing to deploy AI for
26
THE ALLEYNIAN 712
Made with FlippingBook - Online magazine maker