It’s natural for humans to dream beyond what is currently possible to imagine a future in which impossibilities are every day realities. In fact, this is typically the beginning stage of most innovations. Wilbur and Orville first had to imagine what it would be like for a human to fly before they could set about designing a solution. The limitless potential of what we can think up allows us to aim high and try to achieve the never-been-done.
On the other hand, this creative freedom can set expectations that are a little skewed or just too far off to be real. According to the Back to the Future franchise, we should have had flying cars three years ago. Similarly, the movie 2001: A Space Odyssey, gave the mass market a view of a fully autonomous, albeit intimidating, artificial intelligence (AI) in the form of HAL 9000. Coming out in 1968, this movie began to set the narrative for how humans should feel about an AI that has the same intellectual capabilities as a human, otherwise called Artificial General Intelligence (AGI).
Artificial General Intelligence is a machine with the equivalent intellectual ability of a human
To recalibrate our expectations, no one has yet developed a true AGI that mirrors and mimics a human’s full ability. Even if the trepidation around general intelligence began with a movie concept, many pioneers in the technology space have expressed concern about actually developing one. Some specific machine learning use cases exist, such as being able to recognize abnormal credit behavior to detect fraud, or even search engine algorithms to show you more relevant products and information. However, we have used and depended on Artificial Narrow Intelligence (ANI) for years already.
Artificial Narrow Intelligence performs a limited set of functions and cannot learn new tasks by itself
ANI, sometimes called “weak AI”, can be found almost everywhere in the modern world, although some instances are more complex than others. Have you ever played computer chess or minecraft—that’s ANI. A more sophisticated example is Apple’s Siri or Google Home, infused with a wealth of data so they can answer thousands of questions. But there’s still a limit to what they can do.
One of the greatest benefits to ANI is simply the time and effort saved on otherwise tedious tasks. Because an ANI can be built for a sole purpose, an ANI can view, organize, and find patterns in enormous data sets much faster than a human. This allows the human involved to focus on the creative application of the data, something an ANI cannot do.
You may think automating these repetitive administrative tasks are small efforts, but they can add up to be a tremendous improvement to productivity. Have you ever watched a movie or show that was in the “Recommend for you” section of Hulu or Netflix? Or asked Siri for directions? All examples of ANI at work.
In a business setting, ANI is being used review and audit thousands of contracts, record audio calls and make them searchable, and even train healthcare professionals in medical software.
From the few examples I’ve covered here, it’s clear that ANI can be utilized to deliver affordable solutions to customers across a variety of environments, even in secure environments in the federal sector. Our own narrow AI bot, Shai, is currently deployed to help with highly-repetitive tasks such as employee onboard and off boarding, report ingestion, data migration and more. While we may be many decades away from reaching general intelligence, we are already seeing the benefits of narrow AI throughout the enterprise.