The Generative AI Revolution Opportunities, Shocks, and Risks

In a DeepMind study, a range of LLMs were trained in order to understand how performance on a vast array of metrics from coding challenges to mathematical puzzles improve as the size of an AI model increases.12 The size of an AI model is usually measured in two quantities; the amount of parameters used by the model and the number of tokens or the quantity of data used. [...] If that occurs then in the optimistic modelled scenario the productivity benefits from AI alone could lead to greater productivity growth in the UK over the next decade than total productivity growth combined over the previous five years.24 This forecast therefore underscores the transformative power of generative AI in the context of the UK's productivity puzzle. [...] Part of this is due to a lack of access to compute by researchers which the government is looking to remediate with the establishment of the AI Research Resource as announced in the Budget. [...] The Industrial Revolution’s mechanisation of labour gave rise to the Luddites, the introduction of the assembly line in the early 20th century caused clashes with unions, and the introduction of computer technology in the 1960s and 1970s was feared to be the end of white collar work. [...] As part of the UK Science and Technology Framework announced by the Prime Minister and the UK Science and Technology Framework in March, £8 million was committed to encouraging the next generation of AI leaders around the world to do their research in the UK.