Will AI even the socio-economic playing field globally, or will it cause further divides? While we may not know the answer to this question for years to come, a recent study released by Accenture points to increased productivity in countries that have the infrastructure in place to take advantage of AI’s economic boosting powers.
TechEmergence had an opportunity to sit down with Accenture CTO Paul Daugherty and Global Technology R&D Lead Marc Carrel-Billiard, after the AI Summit in San Francisco this past September, to discuss the study outcomes and AI’s economic impact across countries.
The study, released this past summer and titled “Why Artificial Intelligence is the Future of Growth”, was a joint effort between Accenture and Frontier Research. According to Daugherty, it was inspired by one seemingly simple question: “How will AI impact business and what may be its impact on the economy?””
Together, the two organizations set out to study how AI might drive economic improvements by 2035, with a narrowed focus on 12 countries that are responsible for more than 50 percent of the world’s economic output.
AI has the potential to surmount many of the current hindrances to a country’s GDP and total factor productivity (TFP) i.e. how well an economy uses its existing people and capital, by changing the way that industries function, from operational to managerial levels. AI is unique as a “capital-labor hybrid”, meaning it can not only make workers more efficient, but also take over automated tasks and some higher-level responsibilities (emerging virtual assistants and other intelligent machines are early examples), as well as inspire new ideas and help revolutionize infrastructure.
The study’s hallmark finding was that AI has the potential to increase productivity by an average 40 percent across the studied economies, a percentage that assumes AI is part of TFP. In other words, AI will not be just an external influence, like the railroad or desktop computers, but has real potential to be an actual player in moving the economy forward.
Competitive Advantages in the Future Economic Landscape
The study focused on a country’s “national absorptive capacity” (NAC), or how well a country is prepared to absorb and spread innovation. As Daugherty describes the approach, “The kinds of things you need to look at are – what’s the structure of their economy, what kinds of industry do they have, what’s their capacity of research and development, what kind of ecosystems do they have around innovation…you look at the underlying economic data, dig a little deeper and say...how can AI technology change productivity and how will that drive underlying economic improvements.”
While the study finds that AI is poised to double the size of all 12 economies surveyed in half the time (compared to baseline growth) by 2035, this growth is much more in some cases; Japan, for example, could more than triple its gross value added (GVA) growth in 20 years.
Global advantages - like cheaper computing power and increased access to big data - play a hand in AI preparedness, but there are an array of other factors that can affect a country’s readiness to leverage the technology. Geography gives some countries an advantage. “In B2C models, countries like the U.S. and China have an advantage because the size of the consumer populations means you can bring a new business model to scale quickly,” said Daugherty.
A variety of other factors can affect NAC, including technological maturity and public investment; business climate (the US’s strong entrepreneurial climate is a big boost); the strength of particular industries and occupations (Great Britain’s strong aerospace and pharmaceutical industries are primed to better leverage AI in these domains, for example); reliableness of regulatory frameworks; and research and development capacity (Japan has a strong history in this arena, particularly in robotics).
Artificial Intelligence at Work
Daugherty describes an example of an intelligent augmentation project that Accenture undertook with a large international manufacturer: “We looked at their assembly line workers and developed a solution that equipped them with augmented reality headsets that had laser-guided precision assembly capabilities, combined with machine learning capabilities that learned about their performance and best practice on how to do their job, and this technology allowed these workers to do broader sets of work, to increase their capabilities faster.”
This is one of many, and potentially countless examples, of how flexible AI systems can play a role in a broad range of industries and organizations worldwide. Carrel-Billiard also gave an example of a machine learning system being used in novel ways to help strengthen local economies at present: “Look at a platform like Google Translate, it’s an amazing platform...you can leverage and integrate refugees coming to countries to learn about the language...they get integrated faster, they get a job faster – nobody thought about that, but people are using it this way…I think we’re going to see many examples of this.”
Smart systems can be combined to do a variety of tasks on different levels, with capabilities of comprehension (NLP and inference engines are an example), sensing (computer vision and audio processing have made unprecedented progress in the past year) and following through on actions (predictive analytics engines and auto-pilot software in autonomous vehicles are just the surface of the future). As emphasized by Daugherty and other researchers and executives in the field, a central component of “true” AI systems is the ability to learn and adapt over time, with repeatability of these systems at scale, processes that are being explored through fields like reinforcement and continual learning.
Reimagining Business and Industry for the 21st Century
According to Accenture, one of the great challenges for industry is the capacity for businesses to reimagine their business processes. “By combining IoT, machine learning…all these different things, you’re going to be able to change the way we build a product and manufacture a product…the flexibility of these products will not come from the hardware, but from the software,” said Carrel-Billiard.
The key to succeeding in a new era of AI influx is to create a different formula for business, said Daugherty: “Think about each of your business processes, look at the data that’s available and new data that you might acquire, and think about the way your employees and consumers work with systems and machines.” For example, the Accenture study points out that technological education currently goes in one direction i.e. people learn how to use machines, but this will change over time as the learning stream becomes more of a two-way system. In the next decade, customer service reps may act as role models for virtual assistants and vice versa.
How this technology applies in specific industries is still being developed. “We’re at the early stage…the challenge is finding the patterns, and that’s something we’re very focused on…you have to apply patterns across industry,” said Daugherty. “We applied AI in the drug discovery and clinical trials process and pharmaceuticals, and what can you learn about that might apply in the product development process for a consumer goods company, and you have to look at how you can learn from one application and apply it to another.”
Though the report emphasizes an increase in growth and productivity, Daugherty acknowledges that the proliferation of AI - as with any new technology - will result in dislocations in the workforce, perhaps to a greater extent than previously seen, though there are likely to be new types of jobs created as well. “I think there’s an obligation among business and government educational institutions working together to make sure we’re preparing the current workforce for the transition that’s coming and preparing the workforce that’s coming for the new transitions,” said Daugherty.
A push in preparation means acting now to think through the need for new regulation and laws in business and government, and we’re now beginning to see the government respond to innovations in autonomous vehicle and drone development. The Accenture study suggests that AI could also one day be a part of this process, helping to create self-improving regulations to keep up with light-speed advances in technology.