The Future of Leadership/Artificial Intelligence in Organizations

Introduction

Once confined to the realm of science fiction, Artificial Intelligence (AI) has seen significant gains in development and practical application in the previous couple of decades. The recent fame of OpenAI’s ChatGPT and DALL-E 2 has demonstrated AI’s increasing sophistication. As the technology matures, the projected economic impact of AI cannot be understated. Like steam power, electricity, and computers before it, the advent of AI is expected to become a new general-purpose technology (GPT) – a technology with a wide scope for application, growth, and economic impact – forming the basis of the fourth industrial revolution. Even now, businesses and institutions are learning to harness AI to automate tasks previously only performed by human workers. For example, AI is being used to automate data entry and processing, analyze medical images, and automate customer service. The benefits of integrating AI into business and industry is expected to be significant as AI is used to cut costs, enable a more creative workforce, gain insight into hidden trends, and manage tasks and workflows. It is this concern for the integration of AI and business that is the focus of this article. This article seeks to illuminate AI’s strengths and weaknesses, outline where AI is useful in business, and explore the technology’s disruptive impact in the future of work and leadership.

Background

Artificial intelligence, also known as machine intelligence, is a broad term used to describe a group of technologies capable of performing tasks that approximate human intelligence. AI technologies comprise certain types of hardware and software that include computer vision, machine learning, speech recognition and synthesis, and natural language understanding and processing. Together, these technologies synthesize information, recognize patterns, learn from experience, solve problems, and make decisions.

Artificial intelligence works by processing large datasets and adapting algorithms to find new patterns to solve a problem. This means that AI is iterative, it continually improves by running simulations of a problem, measuring its fitness in solving the problem, and exploiting new patterns to improve efficiency in subsequent generations.

Strengths

In general, AI is very good at performing and automating routine tasks that have clearly defined objectives. Such tasks include the reproduction and synthesis of information using existing inputs, computing data to find new patterns, and formulating recommendations. These tasks can be grouped into four broad categories: detection and classification, pattern recognition, prediction, and decision making.

Weaknesses

While AI performs exceptionally well at some tasks, it is constrained by some significant limitations. First, AI cannot produce novel outcomes. Because AI performs tasks within the bounds of predetermined rules and standardized procedures, AI cannot produce any outputs that do not originate from inputs of existing data. Second, AI has trouble guaranteeing quality. Artificial intelligence outputs are dependent on the amount and variety of information fed to it during the training process. However, even large volumes of information do not guarantee optimal or accurate outcomes. This is due to AI’s weakness with generating complexity. Third, AI is limited by significant resource costs. AI requires significant computational processing power, data storage space, large volumes of curated and codified data, and financial investment.

Implications for Leaders

As more industries seek to explore the implementation of AI, some leaders may question whether AI will be beneficial in their business. Though AI has the potential to be very useful for some applications, particularly where companies have access to large datasets, the return on investment of the technology deployed in any business is not always certain to be positive. Before implementing artificial intelligence in a business, leaders are cautioned to observe the technology's two limitations: cost of investment and time before return on investment. First, AI costs can vary widely depending on the configuration. In addition to the AI software, storage space and processing infrastructure will need to be purchased. If the AI is being managed in-house, data scientists will also need to be hired. Costs for AI implementation can total in the low thousands to the millions of dollars. Second, AI development and implementation require a significant investment of time before the technology becomes profitable. Deployment times of AI can take as much as a year.

Despite these limitations, AI can be leveraged to solve problems and create solutions regardless of the size of your business. The trick to implementing AI is knowing the right circumstances and situations to use it. As a guide, before implementing AI it is important to have a clearly defined problem for AI to solve. Next, data are needed to solve the problem, the more the better. Likewise, you need to determine whether the necessary data exists to be fed into the system. Finally, the right personnel are needed to provide oversight for the project including leadership and data scientists to train the AI.