Brent Orrell and David Veldran
In an era defined by rapid technological advancements, artificial intelligence has emerged as a so-called generalpurpose technology, akin to the steam engine, electricity, and the transistor, with the potential to reshape all aspects of our economy and lives. Much of AI’s transformative potential is due to recent developments in the field—particularly the rise of generative AI, which can create novel text, images, and audio. According to a recent report published by McKinsey & Company, this new wave of technology could add $2.6–$4.4 trillion annually to the global economy, with an outsized impact in industries such as banking, software and tech, and the life sciences (Chui, Hazan, et al. 2023). According to Goldman Sachs (2023), AI could raise the global gross domestic product (GDP) by 7 percent over 10 years.
Changes of this scale in growth and GDP will likely affect businesses and workers profoundly. The question is what those changes will be and how we can prepare for them. In this report, we review over a decade of research on AI’s potential and actual impacts on employment trends and demand for skills in the labor market. We then explore this research’s implications for skill development and worker training and offer recommendations for workers and policymakers.
As we stand on the threshold of an AI-driven economy, the future of work is at stake. Will work still be an important human activity, or will human labor be rendered surplus to requirements? While machine-based intelligence has loomed over human imagination for centuries,1 the modern idea of AI emerged in themid-20th century with the development of digital computers. As shown in Figures 1 and 2, while references to artificial intelligence in popular literature rose sharply in the late 20th century, peaking between 1985 and 1990, mentions of the term in academic literature remained relatively low until 2010. As the topic gained ground in scientific circles, references to it in popular literature also began to climb.
AI’s growing role in the public consciousness coincided with revolutionary change in fields such as image recognition and natural language processing, setting the stage for AI’s prominence in public discourse today (Roser 2022).
Between 2010 and today, AI development can be broken into three major periods. From 2010 to 2016, breakthroughs in neural-network design transformed AI. Powered by these new systems, AI made rapid gains in image recognition and deep learning, especially through models such as AlexNet and AlphaGo, a program that shocked scientists and other observers by defeating a world champion Go player (Albrecht 2023; Google DeepMind n.d.).
AI’s advances in 2016–22 built on earlier development with the introduction of highly influential neural-network models including Google’s Bidirectional Encoder Representations from Transformers and OpenAI’s GPT (Albrecht 2023). These models set new benchmarks in natural language processing, helping rapidly improve AI reading comprehension and language understanding.
The third era, from around 2022 to the present, has featured the maturation of generative AI and the proliferation of generative AI tools for broad public use. In this brief period, tools such as ChatGPT, Bard, Claude, DALL·E, Midjourney, and a host of other tools have become widely available, spurring debate about the impacts of AI on work (Chui, Hazan, et al. 2023). Below, we examine the key research voices and trends in each of the three major periods of development outlined above.
No comments:
Post a Comment