The term artificial intelligence (AI) invokes images of robot uprisings, space missions to galaxies far, far away, and lab-created clones that make humans immortal. For years, thought-provoking talks by professors have entertained the notion of whether AI is—or ever will be—self-aware. The more adventurous among us may be drawn toward theosophical discussions on creationism or debates about the realities and influences of the quantum world.
Current thinking about AI may border on science vision (if not science fiction or philosophy)—perhaps for a good reason. Technologies once imagined only on the movie screen now bring convenience and value to our daily lives. Some examples include gestural interfaces, machine-aided purchases, facial recognition, autonomous cars, miniature drones, ubiquitous advertising, and electronic surveillance. Machines are now making predictions on trading stocks, customer purchases, traffic flows, and crime—much as we saw in the 2002 movie “Minority Report.”
From movie screen to real-world applications
Technology leaders have placed big bets on technologies such as brain-computer interfaces, AI in medicine, and deep learning and machine learning tools. AI is expected to lead the new economy, which is becoming known as the Fourth Industrial Revolution or the Second Machine Age. AI is at the forefront of business innovation, along with emerging technologies such as robotics, the Internet of Things, 3D printing, quantum computing and nanotechnology.
Companies are still deciding how AI can be designed to fit into their processes. However, burning questions persist around whether self-learning machines will replace or assist humans in white-collar and blue-collar jobs:
- Can machines learn common sense and empathy?
- Who owns the insights that are generated by AI technology, and who owns the responsibility for an erroneous decision made by a machine?
- Can you teach a machine how to make a decision when dealing with an ethical dilemma?
While these concerns still require much deliberation, most industries understand that the application of AI in businesses brings immense potential. Currently, the top 10 use cases for the technology are data security, personal privacy, financial trading, healthcare, marketing personalization, fraud detection, recommendations, online search, natural language processing (NLP), and smart cars.
Considering how quickly these new technologies are adopted and adapted to new use cases, it is only a matter of time before we start seeing AI capabilities become a part of the fabric of normal business processes. While routine transactions have already been automated, many companies that are higher on the learning curve use predictive and prescriptive analytics to guide their operations.
In the supply chain management function, people talk about degrees of autonomy in the planning process. From use of historical data for planning, it goes through use of automation that can be overridden and ends at nonoptional automation, where planners cannot review the recommendations of the algorithms. The algorithmic supply chain requires organizational maturity and cultural readiness to embed and regularly rely on systems. The concept of an intelligent supply chain goes a step further by incorporating self-learning capabilities of the machine to make better supply-chain decisions.
An opportunity to “learn” and improve–without disruption
Common wisdom tells us that organisations compete on the strength of their supply chain ecosystems. Future organisations would compete on the strength of intelligence embedded in their systems. Ultimately, the winner will be the supply chain that learns most quickly with greatest precision.
At a fundamental level, machine-learning algorithms are a teaching set of data. The machine then answers a question by adding every possible correct or incorrect answer to the teaching set. The algorithm keeps getting better and smarter over time.
Organisations learn in a similar fashion: Every organisation has its own embedded intelligence, which manifests itself through the behavior of its managers and their response to the environment. Supply-chain managers use it to review and modify machine-generated forecasts, production plans, or procurement plans.
Putting a self-learning loop into the system will allow a machine to analyse, for example, why a manual override was made to its recommendation, and it can then check for it during the next cycle. This capability is helpful with managing transactions such as fixing incorrect settings, changing norms, or addressing evolving market dynamics. Over a period of time, machines would learn how managers prioritize their plans based on emerging business scenarios, not just optimization algorithms.
For more on how advanced technology is transforming traditional business models, see Are You Joining The Machine Learning Revolution?