In industries as diverse as finance and medicine, artificial intelligence is now performing work that was previously limited to humans. Autonomous vehicles, smart robotics, explainable AI, language processing, and, most importantly, machine learning promise to transform not only the unglamorous work of inventory management and supply chains, but also the critical front office operations. Along the way, it will shake up old stodgy industries, redefine the nature of work, and create entirely new sources of value.

How Autonomous Decision Making Works

Computers excel at tasks that require understanding a few basic unambiguous rules, but it is poorly suited to more abstract tasks. Machine learning, by contrast, enables a computer to learn from its decisions by crunching vast quantities of data and then extracting rules and patterns from the noise. Most importantly, it is self-teaching. The more data it is fed, the more it will learn. That will enable companies to make entirely new breakthroughs, from automated financial decisions to self-driving cars to personalized medicine.

Artificial intelligence depends on two important trends. First is the amount of data available. Sensors, smart devices, social networks, and individual users are producing approximately 2.2 exabytes, or 2.2 billion gigabytes, of data every day. It would take a standard broadband connection millions of years to sift through it all. The second important trend is the emergence of artificial neural networks, which mimic (in a relatively basic manner) how biological neurons behave. These trends are magnified by the tendency of computing power to double every few years.

The Economic Impact of Artificial Intelligence

An analysis by McKinsey, a consulting group, estimates that the total worldwide investment in artificial intelligence – which encompasses acquisitions, internal research and development, etc – was between $26 and $29 billion in 2016. Many of these investments are still in the experimental phase and may take years to reach the market.

As a result, analysts are divided on the future impact of autonomous decision making AI. Some analysts are more sanguine about the potential of an AI revolution, forecasting that the economic impact could exceed $100 billion by 2025. Others take a more pessimistic view, predicting less than $1 billion.

Nevertheless, as McKinsey points out, the most likely outcome is that artificial intelligence will shape all aspects of society, including governments, corporations, and social institutions. Information will be treated as a vital economic asset, like energy and capital, while delivering significant competitive advantages for firms that are fully geared to it. Artificial intelligence will unlock the ability to:

  • Accurately project inventories and optimize supply chains
  • Anticipate and react to sales trends
  • Minimize waste
  • Predict the success or failure of a prototype before releasing it to the market
  • Deliver more efficient product designs
  • Compress the design process down to a few weeks
  • Bring products to market faster
  • Personalize consumer experiences
  • Better assess the effects of an advertising campaign

The application to manufacturing, service, and technology industries is obvious. When forecasting inventory levels, for instance, artificial intelligence can incorporate historical sales data, supply chain setups, local weather forecasts, and up-to-the-minute price fluctuations. McKinsey estimates that artificial intelligence can reduce costs related to transportation, warehousing, and supply chain administration by as much as 40 percent. Forecasting errors will likely fall by 30 to 50 percent.

How AI Will Change Business

As artificial intelligence takes on more roles throughout the economy, some businesses will be largely automated. This could presage massive job losses, but the most likely outcome is that it will complement, rather than entirely replace, human work. Humans can do things AI that is not good at, including the ability to think abstractly about a problem. AI can do things that humans are not good at, such as spotting errors and reducing material losses, thereby radically improving productivity and minimizing downtime. A collaborative culture between humans and AI will likely emerge, and explainable AI – essentially, the ability of AI to explain its rationale and decision making – will be at the center of it.

Leading tech firms and AI companies will develop most of their AI capabilities in-house. Google’s parent company, Alphabet, has spent billions of dollars acquiring the necessary expertise to become a world leader in AI technology (for example, Alphabet’s acquisition of the British AI company DeepMind). To build up their own data ecosystem, however, the vast majority of corporations will need to partner with or acquire expertise from other leading AI firms.

As AI technology expands in scope and importance, companies will not only need to build the infrastructure necessary to efficiently handle large volumes of data, they will need to acquire the right workforce skill, advanced analytics tools, and corporate structure to optimizing AI performance. Once the AI is producing results, companies will need to incorporate AI insights into the workflow at every level of their organization.