It’s an exciting time, and there is a lot of potential for new technologies to change the ways that we live, and the ways that we do business.
However, sometimes the promotional language doesn’t match the results that you see from a new advancement in IT. Experts (including those at Gartner) talk about a “hype cycle” for new technologies that affects how they are perceived, and how they are used, when they’re brand new.
AI is not immune, and it’s undergoing its own hype cycle right now. These are some of the things that people fail to take into account when reckoning an accurate potential of artificial intelligence.
AI in the Real World
Many AI entities are very good at taking in input, and spitting out results based on language models, but they may not be able to deal with real world decisions and really analyze their surroundings in detail.
For example, you might see AI tools that can accurately predict the best response to a question, or sven ‘see’ around them with sensor based systems and convolutional neural networks. But they may have gaps in their ability to really discern their environment. They might not recognize common objects, or be able to identify what they see fully. These gaps can be dangerous, and even fatal, as in some of the cases around technologies like a certain self-driving autopilot system in its early iterations.
So this is something to keep in mind as we move forward.
Machine Learning and Artificial Intelligence
Others looking at the relentless hype around new applications point out that some of the cases that we refer to as AI are actually machine learning programs, something, for instance, that is predictive but not really cognitive in a human way.
In other words, AI is kind of a vague term to talk about systems that might be able to do certain tasks in the ways that we do, but are not ‘thinking’ in the ways we suppose that they are. We see them as ‘like us’, but in reality, they’re much different. It can make a lot of sense to think about how Ai entities and humans see things differently, recognize different concepts, and work differently, even though they may be chasing the same ultimate answers.
So with that in mind, rather than just saying that AI ‘does’ this or that, we should be assessing how it does it. Does it do it based on heuristic data and guesses? Does it really have the ability to determine a situation and the best response, or is it just fooling us by passing any number of Turing tests?
When we start to answer these questions, we’re cutting through the hype, and looking at the real face of artificial intelligence as it is today.
Companies Talking About AI
Then there’s the phenomenon of hype where companies are talking about everything that they’re going to do with AI…but when you look around the industry, not much is being done with AI yet.
The numbers can be confusing, if you’re going by the number of people who are mentioning AI in corporate literature or anywhere else. Does that actually translate into action?
You have to actually look at where the technology is being applied to get an accurate picture of how it’s used. You also have to remember that integration and staff training are critical: joining a technology to an existing business process the right way can make the difference between this innovation being a help or a hindrance to the company.
Recognize AI Deficits
In many cases, AI hallucinates. It makes errors. It’s not all powerful or omniscient. But it fools people into thinking that they’re dealing with something infallible – until, that is, the AI makes a mistake.
This is part of the ethical AI idea, where we develop clear ideas about how the system makes determinations, and put that data out there for everyone to see. We want to be sure that we see whether the AI is on task or not, and whether its products are true. That’s something that users ignore at their peril.
How to Find Real Value with AI
So how do companies move forward in using AI the right way?
First, you want to identify the company’s goals and objectives, and make sure that AI can address them. What sort of workloads and processes are you going to automate, and how? What is that going to do for the business process?
Then, too, companies can leverage the predictive AI and the data to create business insights and automations and other improvements. For example, they may be able to move a business process closer to a desired state, or introduce key efficiencies, based on what the AI can actually do on a consistent basis.
Data governance is another big issue. Companies have to treat the data well and use appropriate methods to drive their systems. That means thinking about data ownership, intellectual property, etc. – with those things in hand, companies can move closer to real implementation. That’s going to move the ball forward.
Those who can avoid the normal pitfalls of the AI hype cycle are much more likely to get further in an era where these technologies are being blended into our lives at a startling pace.
