It’s hard to escape the footprint of artificial intelligence. Though once thought of in futuresque, sci-fi terms, artificial intelligence (AI) in its various forms has proliferated everyday life both at home and in the enterprise. According to a report from Gartner, the number of enterprises using AI has tripled in the last four years, meaning that from Amazon product recommendations to voice-controlled light switches to automated functions that improve key pain points and boost productivity at work, AI is now everywhere.
But what does AI actually mean and what do you need to know about it now? Keep reading to learn the basics.
What Is AI?
According to the Oxford Dictionaries, artificial intelligence is “the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.” In other words, AI is the creation of software capable of performing tasks that once only humans could do. This can include anything and everything from scanning documents for certain types of data to playing chess to driving cars. Making the concept even broader, AI consists of both the underlying processes required to deliver such functions and the computers and robots capable of fulfilling the tasks in place of humans.
Types of AI
AI and its subsets are sometimes referred to interchangeably, but in fact, there are numerous different types of AI and various applications for each.
At the broadest level, AI can be divided into Artificial Narrow Intelligence (ANI), or “weak AI,” Artificial General Intelligence (AGI), or “deep AI,” and Artificial Super Intelligence (ASI). Currently, broad AI capabilities are classified as ANI, which means computers’ ability to perform human actions is isolated to a narrow set of predefined parameters. This does not mean they are necessarily very limited or unsophisticated – for example, chatbots and self-driving cars are capable of responding to a broad variety of inputs – but it does mean that most AI actions are still fairly specific and narrow.
AGI is the next level of progression where machines are not only capable of performing tasks as humans can, but of performing them so well that their actions are indistinguishable from humans’. This includes being able to make decisions under uncertain conditions, employing reasoning and strategy, planning, learning and being capable of communicating in plain language. The last subset, ASI, refers to the – as yet unrealized – highest potential of AI, in which machine behavior not only surpasses human capabilities but also is capable of rapid ongoing learning and improvement.
The Functionalities of AI
In addition to its overarching capabilities and potential capabilities, AI also offers different subsets of functionalities. These include:
Machine learning – where a computer uses algorithms and statistical models to teach itself how to perform functions, with each new interaction helping to refine outputs;
Robotic process automation – in which a computer or robot uses rules-based code to automate ongoing repetitive tasks, such as data entry and assembly lines
Natural language processing – the processing of natural language such as speech or emails to enable computer-human interactions
Speech recognition – the translation of spoken language into text, which can be then be processed by computers
Some of these functionalities may work together or go hand in hand. For example, devices capable of speech recognition will also rely on natural language processing to convert human voice commands into ones a computer can understand. Machine learning can use the results of various artificial intelligence interactions, whether automation or voice commands, to improve future performance.
Key Benefits of AI
While AI-enabled tasks are still somewhat narrow, artificial intelligence is already proving to deliver numerous benefits, helping humans to achieve better, faster and easier results both at home and in business.
According to one report, automation at work could save businesses nearly $5-million annually. Where do those savings come from? According to the report, 53 percent of employees say they could save two hours per day by skipping tasks that could instead be automated, netting 240 extra hours per year per employee. Seventy-eight percent of business leaders say they could save three hours a day on automatable tasks. Shifting office jobs such as data entry, tagging documents with metadata and other mundane tasks to automated workloads saves time that can be reinvested into higher-value work, and also potentially may increase satisfaction on the job due to the elimination of drudgery. In contract management, for example, employing automated tagging for on-boarding new agreements could free up time for better contract post-mortems, allowing contract managers to identify language and terms that have resulted in the most successful implementations and using those learnings to create better standard agreements.
Because computers are capable of processing high volumes of information and are less vulnerable to errors the same way that humans are, AI can also help both businesses and people to make better, faster decisions that take a more considerable amount of data into account.
We’re still just beginning to unlock the potential of artificial intelligence in our work and personal lives. Keeping on top of the advances is a good way to leverage new capabilities as they become possible.