The Limitations of Artificial Intelligence

In recent years, there has been remarkable progress in artificial intelligence (AI). The ability to incorporate human-like intention and adaptability into AI algorithms has opened up enormous opportunities for the application of the technology to enhance human performance and augment people’s capabilities. At the same time, these advances raise concerns that the technology may lead to doomsday scenarios sensationalized by movie-makers—or at least threaten basic values and endanger the essence of humanity.

The term “AI” has no uniformly accepted definition, but it is generally defined as machines that exhibit characteristics of human intelligence in tasks such as reasoning, decision-making and control. AI has been developed to complete tasks that humans find hard or impossible, including recognition, information retrieval and manipulation, design and operation automation, and intelligent personal assistants like Siri, Alexa, and Cortana.

Almost all computers today use some form of AI. It’s built into a variety of technologies, from smart phones and computer search engines to autonomous cars and factory robots. However, the full vision of “general” AI that you see in sci-fi films is not yet possible outside the laboratory.

The limitations of current AI can be seen in the way that it learns. Most AI systems are taught to perform specific tasks, such as playing chess or finding the best product to recommend online. They learn by analyzing huge training datasets, which are data sets that teach algorithms how to make decisions, and then they adapt as they take on new input.

While they have a long way to go before they can match the full range of cognitive skills that humans possess, the development of AI systems has brought many benefits to society and business. These include better recommendations, faster service, more precise decision-making, and improved automation.

Businesses that adopt and integrate AI are growing at a much faster rate than those that do not. However, many organizations are still dipping their toe into the AI pool instead of jumping in headfirst. The companies that take the lead in leveraging AI can gain a significant advantage over their competitors and transform their operations from siloed to interdisciplinary work.

For example, banks feed AI systems large amounts of training data about fraudulent and non-fraudulent transactions to learn how to recognize them. This reduces the number of manual, time-consuming processes for detecting fraud and saves on processing costs. Likewise, many e-commerce websites now employ AI chatbots to serve customers, eliminating the need for employees during peak hours and freeing up resources for other functions. This has also led to more personalized and relevant customer service. For more, see McKinsey’s article on How to harness the power of AI. artificial intelligence