One of the most talked-about technologies today is artificial intelligence (AI). Although much of the discussion is still theoretical, the technology is developing rapidly, and it has already begun to transform how some industries operate. In the past, capital investment and labour have been used to increase production and profitability. These tools have become significantly less effective over time, raising the question of whether our economic system should change, and how. Some believe that AI will be the innovation which kicks off a new industrial revolution. A forecast by Accenture suggests that AI could boost overall profitability by 38%, adding US$14 trillion in gross value added by 2025. The rise in profitability could be as much as 50% ten years later. And, although predictions are never certain, many business leaders seem to agree. The BCG and MIT Sloan Management Review has reported that 85% of global executives believe AI will give them an advantage in the marketplace. In this article, we’ll discuss exactly what AI is, and what possibilities it offers. We’ll also review who stands to benefit, how to prepare your business, and some of the emerging ethical and legal issues around AI.
What is AI?
The essential goal of AI research is to create devices which can think and adapt. Traditional automation meant building a machine which could complete specific tasks. What makes AI different is that the machines and software involved are capable of learning by themselves. They have the potential not just to automate processes, but to improve, develop and even redesign them.
Insight: AI business boom
- AI will be responsible for 85% of customer service interactions by 2020, driving up to US$33 trillion of annual economic growth.
- Information and communications, manufacturing, and financial services will all see the biggest real-terms growth as AI becomes more integrated into their day-to-day processes.
- Education, construction, and accommodation and food services will earn the greatest relative profit increases.
Most researchers talk about two different kinds of AI: narrow and general, also known as weak and strong. Weak AI is designed to complete a single task, such as automating a factory process, or understanding and responding to voice commands. Some tasks which used to be considered weak AI – such as computers recognizing and recording text – are now so commonplace that they are not really a part of AI research. Strong, or general, AI would be able to learn and think in a holistic way. It would not be restricted to single tasks. There is a wide range of estimates about how long it will take to develop this technology: some researchers expect general AI to exist by 2060, while others predict that it will take hundreds of years.
How can AI boost profits? According to a recent Accenture report, AI can transform industries in three ways. There’s intelligent automation, as we’ve discussed briefly above. Better automation frees up time and money – this is described as labour and capital augmentation. And finally, AI makes innovation faster and easier to implement. Automation has long been used as a way to increase profits. It can save on time and labour, helping to reduce overheads and increase productivity. Intelligent automation builds on these benefits and takes them even further. AI makes it possible to automate more and more complex processes – even in tasks which we once thought could only be performed by humans, such as teaching or customer service. It also offers the opportunity to improve automation where it is already in use, by harnessing the power of new AI technology to learn and adapt. Labour and capital augmentation is a more subtle benefit of AI, but it could still have a large impact. More and more tasks can be delegated to intelligent automation, freeing up human workers to work on more interesting and innovative projects – and increase profits overall. Thirdly, the innovative potential of AI could lead to a wave of new products, capabilities and revenue streams. By integrating AI into research and development, all industries have the opportunity to speed up change in how they operate.
Who will the big winners be?
“AI has significant implications for business managers and leaders, too. To succeed in the AI economy, businesses will need a clear, future-proof AI strategy. With new modes of working, new roles in the workplace, and new innovations powered by AI, businesses will need new ways to measure their performance. AI is going to require a fundamental change of mind-set and smart businesses will monitor its progress and think creatively.”
Although applications of AI are almost endless, some industries will reap bigger benefits than others. In addition, some countries will enjoy the advantages of AI more than others. Information and communications, manufacturing, and financial services will all see the biggest real-terms growth as AI becomes more integrated into their day-to-day processes. The benefits of augmented labour and capital, as well as innovation, will be felt strongly in these industries. Education, construction, and accommodation and food services will earn the greatest relative profit increases, as intelligent automation reduces the overheads traditionally required to operate. According to one prediction, AI will be responsible for 85% of customer service interactions by 2020, driving up to US$33 trillion of annual economic growth. The national economies with the most to gain from AI are all in North America or Northern Europe, with the notable exception of Japan. The United States, Finland and Sweden are predicted to have the biggest increases in gross value added and labour productivity with AI.
New ideas, new problems Transparency and data The development of AI goes hand in hand with the revolution in big data. Many AI programmes use a technique called “deep learning”, where they are shown a huge library of data, and taught to recognize and reproduce patterns. For example, some famous AI computers have learnt to play chess by reviewing hundreds of thousands of games. Another AI, known as Aiva, has learnt to compose music after listening to many hours of classical music. Although these might sound like party-piece demonstrations, the deep learning method can be applied in very practical ways. Collecting data about customer service interactions can teach AI how to talk to customers or predict what choices they are likely to make. Crowdsourced data and customer data could be used to super-power deep learning for specific industries and even individual businesses. However, the use of data instantly raises issues of transparency and privacy. Legislation around the world is slowly beginning to catch up to modern technologies – as with the recent introduction of General Data Protection Legislation in the European Union. It remains to be seen whether AI and data usage for AI will be covered by existing laws, or require new legislation. As well as the legal aspects, businesses must also be ready to explain their data usage and AI strategies to customers.
Jobs for people, jobs for AI Many people have raised concerns about AI replacing human workers in the economy. In the past, automation has led to job losses; intelligent automation could lead to many more. However, there are many different opinions about how significant the real impact will be. Some observers point out that after the first Industrial Revolution, new jobs developed which could not have been imagined in the days before factories. Later, the same advances in technology which made widespread automation possible also created new fields, industries and roles for workers. It seems likely that intelligent automation – and the innovation and augmentation which it brings – could deliver new opportunities for human workers. It must be said that not all forecasts are so optimistic. Some well-known figures in the world of technology, such as the computer scientist and philosopher Jaron Lanier, have warned that AI could transform our current model of employment beyond recognition. Some industry leaders and experts have suggested that governments may need to introduce a Universal Basic Income to support people whose roles have been filled by AI. Another suggestion has been Bill Gates’ so-called “robot tax”, a levy on companies using AI which would be used to fund social care.
The sci-fi scenario The final question on many people’s minds is safety. If AI is intelligent, can it be trusted? While science-fiction writers love the idea of a malevolent, all-powerful AI, most researchers think that’s an unlikely scenario. Instead, there may be in a risk from AI which is programmed to do something destructive, or which develops a destructive way to complete a task. For example, some experts have warned of a potential new arms race in “smart weapons”, such as autonomous drones which could use AI. However, other ethical problems in AI could be more subtle and mundane. In a recent experiment, a Google AI chatbot, which was using Twitter to learn about language, began to produce violent and offensive content within 24 hours. (Google has already released its own AI ethics guidelines, and the bot was quickly shut down.) Microsoft has taken the lead in discussing diversity in AI. If AI is taught to recognize faces and voices, it’s important to make sure that all faces and voices are treated equally. The risk of deep learning built on data from people is that it could reproduce biases which people have, such as racism or sexism.
Adapting to the AI reality
New leaders The advent of AI won’t just change life for ordinary workers: it has significant implications for business managers and leaders, too. To succeed in the AI economy, businesses will need a clear, future-proof AI strategy. Senior roles will change to include both managing humans and supervising AI – and this will require a mix of communication and technical skills. Finally, the data needed for effective AI makes a whole new role necessary. Accenture describes this as the “chief data supply officer”, responsible for providing, securing and monitoring data.
New metrics With new modes of working, new roles in the workplace, and new innovations powered by AI, businesses will need new ways to measure their performance. For example, processes operated by intelligent automation will still need to be supervised. Workers who are freed up for different roles by labour augmentation will have different needs in terms of supervision, evaluation, and training and development. Of course, navigating all these complex changes is justified by the potential financial benefits of using AI. Smart businesses will develop metrics to measure just exactly what those benefits are, compared to the profits and growth that would have been achieved without the introduction of AI.
New methods AI is going to require a fundamental change of mind-set. Its power to disrupt traditional ways of working means that we have to look at labour and capital assets with fresh eyes, and be ready to exploit them in unexpected ways. Technologies which allow businesses to collect and manage data on a huge scale have already transformed many industries, enabling them to target strategies and streamline operations. AI can speed up or even automate the process of interpreting data. It has the potential to work alongside humans in analysis and planning. It can also be used to take care of more mechanical tasks, freeing up human brain power and communication skills for where it is really needed. As AI develops, smart businesses will monitor its progress and think creatively about how each innovation could be applied to their own work. They will be prepared to consider original or unexpected changes to their business strategies, and they will be conscious of the balance between AI and human skills.