Artificial Intelligence: Expectations vs. Reality

Let’s face it; we humans can be drawn to exaggerate and over-generalize when it comes to modern technology, especially during its early-adoption, nascent phases. This is understandable, many major tech introductions have the potential to change the way the world works or how we execute specific tasks. For that particular reason, the uprise of not-seen-before technologies can create ripples of attention as the start to shake-up the status quo. Moreover, Sci-Fi movies with not-so-accurate technology themes can feed that disquiet, causing the public to create exaggerated notions of what technologies like artificial intelligence would mean for humanity. Time-traveling liquid metal murder robots from the future may seem too far-fetched, but computers which have all the smart answers and can bring about World peace by exposing crimes before they’ve happened? Yes, much more plausible to the average non-technical person even though this would require a coordinated effort from multiple disparate government agencies who are suddenly magically released from the constraints of fiscal budgeting and the regular popularity contest of elections. These plausible possibilities, in turn, increase the need to define what’s real and what isn’t with modern technologies like AI. 

There are a lot of misconceptions about artificial intelligence. From intelligent robots eventually turning against and enslaving humans in a dystopian societal apocalypse to AI replacing humans in all areas of life. The AI hype is real. In this article, we’ll be discussing some common AI expectations and defining their underlying reality. 

Are we about to face an AI Apocalypse?

From The Matrix to The Terminator, almost all AI-related movies portray a future where mankind is overwhelmed by technology which has turned from friend to a foe whose main, single-purposed goal is our annihilation. The AI turns against the human race or tries to take over the world. This isn’t entirely bad, as long as it is all understood to be as fictional as a world populated by dragons and reincarnated dinosaurs. But this is not often the case; the media influence is, tangibly, stronger than ever, and a large percentage of the public translates this easily-consumed fiction into an absolutely likely reality. In other words, such movies add to how the world sees AI innovations. No matter how close to modern times you foray, you’ll find that the narrative remains the same; machine revolting against humans, often with an underlying moral narrative of “well, we got what we deserved”. Almost none portray a world which results from a more probable extension of the various forms of artificial intelligence that actually exist, with fiction being mainly about the human-like type AI. 

The Royal Society Report explains the implications of exaggerated expectations and fictional narratives. It reads as follows;

Both fictional and many non-fictional narratives focus on issues that form either a very small subset of contemporary AI research or that are decades if not centuries away from becoming a technological reality. This disconnect between the narratives and the reality of the technology can have several major negative consequences. The prevalence of narratives focused on utopian extremes can create expectations that the technology is not (yet) able to fulfill. This, in turn, can contribute to a hype bubble, with developers and communicators potentially feeding into the bubble through over-promising. If such a bubble bursts because the technology was unable to live up to the unrealistic expectations, public confidence in the technology and its advocates could be damaged. The case study on nuclear power illustrates this pattern.

The reality of the AI apocalypse myth is that it refers to a specific type – General AI – an algorithm that can do everything a human can or probably better. We don’t have such algorithms at the moment and, according to experts, we may not be having that any time soon. So, in a nutshell, AI is not taking over the world. Well, not yet anyway.

Is AI after your job?

There’s a popular misconception that AI will take human jobs. This notion is somewhat different from the first in that it is based on some truth. Automation is already replacing the need for humans to perform certain tasks. The effects of globalization and advances in the science of mass organized transportation logistics have shrunk our world. That has grown to mean that ‘smart’ just-in-time manufacturing and delivery technology has impinged on most Western economies where outsourcing of labor, manufacturing and employees backed up by increasingly sophisticated algorithms to ensure a constant supply of raw materials, reductions in expenditure and yes, even computing is trampling those who either did not or could not adapt in time.

Real current AI means shelves are stocked based on the crunching of tangible facts such as units sold but also by more ephemeral points of data drawn from machine learning which has analyzed our buying habits, social media likes, trending topics and the likely type of Winter expected this year. The roles of the finger-in-air buyers and marketers who used a combination of experience and gut instinct have been made to step back while the machines begin to step in and do such a good job that the selections suggested to you are so uncannily good you’d swear that “they are recording our conversations”. No – they’re not, mostly – they’re just peering into almost every other little tidbit of data they can though and there’s enough available that, when coupled with some cleverly trained learning models they can absolutely know what types of apples you like, how many flights will be needed to Kansas in November and exactly when the traffic lights in London’s Oxford Circus will probably fail.

If your job relies on an element of guesswork and a ‘keen eye’ you, and a lot of us, will find that AI will continue to take on human jobs in the coming years. 

When is AI a good thing?

However, the concept we don’t talk about is that artificial intelligence won’t be replacing but will be making jobs easier for most people. Talk about working with AI to get things done faster and more efficiently. This is because there are a lot of jobs where human factors cannot be replaced. In other words, jobs that require human intuition or emotions – AI does a dismal job of this kind of thing. For example, AI in the medical industry cannot totally replace doctors or other practitioners, partly because it’s a job which requires a holistic view of humanity which includes subtle clues like a patient’s general demeanor and other almost subconscious observations which are, thankfully, difficult to quantify and replicate. However, AI can be employed, like any other tool, to help with decision-making, amongst other things and to bolster the diagnostician’s process. 

What fuels this “AI will replace us all” misconception is the notion that algorithms continue to work with the same accuracy once deployed. In reality, this is not the case. Over time, the machine learning model degradation occurs and the model’s accuracy slowly depletes as it takes in new data and patterns. Therefore, continuous monitoring and improvement are required to mitigate model degradation. 

Does My Business Need Artificial Intelligence?

Regardless of the exaggerated expectations and narratives surrounding AI, you’ll agree that it is a great innovation with potential to bring success to not only the business world but to our lives in general. From smaller tasks like sorting a tsunami of social media feeds to maintaining clients’ inventory, AI can increase an organization’s productivity and efficiency. If you’re looking to add AI to your applications in this fast-paced economy, you need technologies that will help you build and market faster. For example, development with Delphi/RAD studio is 5x faster than developing with other languages. 

Click here for or more articles about artificial intelligence on this site.

Why not download a free trial copy of RAD Studio today and experience the power and ease of use for yourself?

What's New for RAD Studio 11