All Resources

Understanding AI for the enterprise




accountancy , AI

Much of the narrative around AI hovers at extremes, ranging from a panacea to an existential threat. Science fiction and popular culture have leaned towards the latter. Awareness of the pop culture coloured threat has led to a misunderstanding if not lack of understanding around the real potential in a range of business contexts.

At Dayshape, we want to share our knowledge of the technology and explain how it can help improve businesses and people’s working lives. But first, we need to unpack what AI is. It’s important to begin a discussion by setting out first principles. By understanding AI, we learn how it can help us.

In truth, it’s complicated

If asked, people tend to have a vague idea about what AI is. Most understand the premise of computers having some form of intelligence, but they can sometimes lack a nuanced understanding of what this means and how it could benefit them, their business, and society as a whole. 

Some people can be put off from engaging with the topic of AI thanks to the role that culture has played in making AI an intimidating proposition. Television shows such as Black Mirror and films like Ex Machina depict a disturbing world where artificial intelligence transcends human intelligence. 

Hyper-intelligent machines can pose ethical questions, and we’re seeing these raised in relation to future technologies, such as autonomous vehicles. For example, the trolley problem asks whether a human would alter the path of a runaway trolley to spare the lives of five people but kill one. How do humans answer this question? How does a machine? 

Thankfully, for our purposes, we need to understand a more basic version of AI, a version already being put to work in the world today. 

Exploring narrow AI for the enterprise

For the purposes of our discussion, let’s focus on “Narrow AI”, which means AI that performs a defined task. Air traffic control systems, driverless cars, and smartphones are all run by narrow AI algorithms. These are designed by humans with computers being trained and put to task. They work for sustained, error-free periods within a system’s parameters, subject to human oversight. 

This definition allows us to focus on a version of AI that’s as easy to comprehend as basic machine learning. It’s a form of regression analysis, another application of statistics, in which AI completes decision-making chains and regressions faster than humans ever could. 

The pace of this progression has been illustrated by victories for narrow AI over humans in chessthe game show Jeopardy! and the Chinese board game Go. Despite these victories, narrow AI still runs into problems. 

Narrow problems 

Problems with narrow AI occur for the same reason that problems occur in enterprise. Narrow AI depends on a human to put it to work, so, therefore, it is still subject to human failings, such as people cutting corners, setting overly ambitious targets, or prioritising incorrectly. 

A reliance on fallible humans means that narrow AI is not a panacea, as it is wholly reliant on being put to task by its master. If that human wrongly defines the task, then it doesn’t matter how long or hard the machine works, it will find a false conclusion. 

To overcome this problem, scientists and researchers are developing Artificial General Intelligence (AGI) or strong AI, which allows for the possibility that machines can design systems of their own. However, this is still some way off, and it still has limitations. 

The most famous criticism of strong AI is the Chinese Room problem, which claims that a computer executing a program can never have a “mind” or “consciousness”, regardless of how human-like the program may make the computer behave. Even strong AI always falls short of being a true substitute for human consciousness.

Using AI to solve business challenges today

Narrow AI can help businesses to solve problems today. By embracing its strengths, we can maximise human brainpower, too. However, it’s important that all AI findings are considered by a human who finds the most appropriate solution for a situation. AI offers options, but a human must work in tandem with the AI to find a conclusion that works best for all participants. 

AI isn’t perfect and many of the perceived benefits are still consigned to the television or movie screen, as are, thankfully, the potential threats. However, the benefits for businesses are endless and there is no better time for business leaders to be engaging with AI from an enterprise perspective than today, exploring gains that can be unlocked over the short and medium-term.

We don’t know exactly what the future holds for AI. But, in 2020, with a basic grasp of what AI is and what it is not, there are real-world opportunities for businesses to explore and take advantage of today. 

Related resources

See all insights