Artificial intelligence (AI) captured the world's attention in 2023 with the emergence of pre-trained models such as GPT, on which the conversational AI system ChatGPT is based. For the first time, we can converse with an entity, however imperfectly, about anything, as we do with other humans. This new capability provided by pre-trained models has created a paradigm shift in AI, transforming it from an application to a general-purpose technology that is configurable to specific uses. Whereas historically an AI model was trained to do one thing well, it is now usable for a variety of tasks such as general conversations; assistance; decision making; and the generation of documents, code, and video--for which it was not explicitly trained. The scientific history of AI provides a backdrop for evaluating and discussing the capabilities and limitations of this new technology, and the challenges that lie ahead.

We investigate the potential implications of large language models (LLMs), such as Generative Pre-trained Transformers (GPTs), on the U.S. labor market, focusing on the increased capabilities arising from LLM-powered software compared to LLMs on their own. Using a new rubric, we assess occupations based on their alignment with LLM capabilities, integrating both human expertise and GPT-4 classifications. Our findings reveal that around 80% of the U.S. workforce could have at least 10% of their work tasks affected by the introduction of LLMs, while approximately 19% of workers may see at least 50% of their tasks impacted. We do not make predictions about the development or adoption timeline of such LLMs. The projected effects span all wage levels, with higher-income jobs potentially facing greater exposure to LLM capabilities and LLM-powered software. Significantly, these impacts are not restricted to industries with higher recent productivity growth. Our analysis suggests that, with access to an LLM, about 15% of all worker tasks in the US could be completed significantly faster at the same level of quality. When incorporating software and tooling built on top of LLMs, this share increases to between 47 and 56% of all tasks. This finding implies that LLM-powered software will have a substantial effect on scaling the economic impacts of the underlying models. We conclude that LLMs such as GPTs exhibit traits of general-purpose technologies, indicating that they could have considerable economic, social, and policy implications.

Major advances in Machine Learning (ML) and Artificial Intelligence (AI) increasingly take the form of developing and releasing general-purpose models. These models are designed to be adapted by other businesses and agencies to perform a particular, domain-specific function. This process has become known as adaptation or fine-tuning. This paper offers a model of the fine-tuning process where a Generalist brings the technological product (here an ML model) to a certain level of performance, and one or more Domain-specialist(s) adapts it for use in a particular domain. Both entities are profit-seeking and incur costs when they invest in the technology, and they must reach a bargaining agreement on how to share the revenue for the technology to reach the market. For a relatively general class of cost and revenue functions, we characterize the conditions under which the fine-tuning game yields a profit-sharing solution. We observe that any potential domain-specialization will either contribute, free-ride, or abstain in their uptake of the technology, and we provide conditions yielding these different strategies. We show how methods based on bargaining solutions and sub-game perfect equilibria provide insights into the strategic behavior of firms in these types of interactions, and we find that profit-sharing can still arise even when one firm has significantly higher costs than another. We also provide methods for identifying Pareto-optimal bargaining arrangements for a general set of utility functions.

In the 1990s, business process reengineering was all the rage: Companies used budding technologies such as enterprise resource planning (ERP) systems and the internet to enact radical changes to broad, end-to-end business processes. Buoyed by reengineering's academic and consulting proponents, companies anticipated transformative changes to broad processes like order-to-cash and conception to commercialization of new products. But while technology did bring major updates, implementations often failed to live up to the sky-high expectations. For example, large-scale ERP systems like SAP or Oracle provided a useful IT backbone to exchange data, yet also created very rigid processes that were hard to change past the IT implementation. Since then, process management typically involved only incremental change to local processes -- Lean and Six Sigma for repetitive processes, and Agile Lean Startup methods for development -- all without any assistance from technology.

In this episode of the McKinsey Global Institute's Forward Thinking podcast, host Michael Chui speaks with Jeffrey Ding, researcher and founder of the ChinAI Newsletter, about information asymmetry in artificial intelligence between China and the West. They cover why data may not be like oil, the Chinese industry adage on products, platforms, and standards, "unsexy AI," and more. An edited transcript of this episode follows. Subscribe to the series on Apple Podcasts, Google Podcasts, Spotify, Stitcher, or wherever you get your podcasts. Anna Bernasek, co-host: Michael, there's a lot of talk right now about artificial intelligence, or AI, and what it means for global competition. I'm really glad we've got a guest today that can talk to us about what's really going on, particularly when it comes to the US and China. It definitely is a fascinating topic--at least, I find it personally. I'm a former AI practitioner and more recently, at the McKinsey Global Institute, have been able to study the impact of AI on business and more broadly. And one of the reasons I'm so excited about today's conversation is because it's with somebody you probably don't know yet but probably should. He's famous in certain corners of the internet but his work, it turns out, is relevant everywhere.

Alan Turing (1912-1954) was one of the first thinkers to take the concept of artificial intelligence ... [ ] seriously. His pioneering work laid the foundation for the fields of digital computing and AI as we know them today. In the most direct sense, artificial intelligence is an engineering challenge. The mathematics underlying today's cutting-edge AI algorithms is complex. The amount of computing resources required to train state-of-the-art AI models is formidable.

Alan Turing (1912-1954) was one of the first thinkers to take the concept of artificial intelligence ... [ ] seriously. His pioneering work laid the foundation for the fields of digital computing and AI as we know them today. In the most direct sense, artificial intelligence is an engineering challenge. The mathematics underlying today's cutting-edge AI algorithms is complex. The amount of computing resources required to train state-of-the-art AI models is formidable.

Alan Turing (1912-1954) was one of the first thinkers to take the concept of artificial intelligence ... [ ] seriously. His pioneering work laid the foundation for the fields of digital computing and AI as we know them today. In the most direct sense, artificial intelligence is an engineering challenge. The mathematics underlying today's cutting-edge AI algorithms is complex. The amount of computing resources required to train state-of-the-art AI models is formidable.

"If delivered, this impact would compare well with that of other general-purpose technologies through history," notes McKinsey. "Consider, for instance, that the introduction of steam engines during the 1800s boosted labor productivity by an estimated 0.3 percent a year, the impact from robots during the 1990s around 0.4 percent, and the spread of IT during the 2000s 0.6 percent." The McKinsey report is based on simulation models of the impact of AI at the country, sector, company and worker levels. It looked at their adoption of five broad categories of AI technologies: computer vision; natural language; virtual assistants, robotic process automation, and advanced machine learning. Data sources included survey data from approximately 3,000 firms in 14 different sectors and economic data from a number of organizations including the United Nations, the World Bank and the World Economic Forum.

It has been 21 years since IBM's Deep Blue supercomputer checkmated chess champion Garry Kasparov, marking a historic moment in the development of artificial intelligence technologies. Since then, artificial intelligence has invaded everyday objects, such as cell phones, cars, fridges, and televisions. But the world economy seems to have little to show for the proliferation of smartness. Among advanced economies, productivity growth is slower now than at any time in the past five decades. National GDPs and standards of living, meanwhile, have been relatively stagnant for years.