Ai Hype. Don't Be Fooled

The Ai Hype Train Has Left the Station

We know we’re in the middle of a hype cycle when every industry has products that use some buzzword, regardless of whether or not they have the technical merit or legitimacy to actually use that buzzword. This is where we currently are with regards to the market for Ai (Artificial Intelligence) products and services.

We hear briefings from vendors pitching technology across a wide range of solution areas, as well as from professional services and government contracting firms with new solution offerings and end users looking for guidance and advice. On the whole, we’ve seen a lot of great, new innovation that’s pushing the industry forward towards more intelligent, autonomous systems capable of addressing more of the challenge areas that have previously not been able to be solved due to extreme complexity or the need for human labor.

However, we’ve also seen more than our fair share of vendors, so-called Ai Demo Days, customer engagements, and service pitches that claim to be Ai-enabled when all they have done is put some thin capability (usually a third-party library or API) that somehow miraculously transforms their automation or analytics or cybesecurity product of yesterday into some new-fangled Ai miracle. We scratch our heads and wonder if this new offering is making anything truly intelligent or is this just the same old product with primarily the same old features sold to the same old customers providing the same old benefits with some new thing added on. Recently, we spent time examining an entire category of product solution that has been known to call itself plainly an automation offering, robotic in fact. This whole category is also currently attempting to rebrand itself as intelligent and Ai-enabled because they’ve added OCR or some other add-on. Baloney. Automation is not intelligence.

What is Automation?

Let’s be clear – automation is not a bad word. In fact, the whole movement of the industrial revolution was to take much of what humanity was doing at the time and automate it so that we could achieve

  • significantly greater productivity
  • quality of life, and
  • transformation of society through technology

The steam engine, train, factory line, and then computers and the internet have truly revolutionized the way we work, live, and exist. However, these are not intelligent technologies. We can’t walk up to a steam engine and ask it to recognize who we are or answer a random question or even learn from its experiences. A web server is just a web server no matter how many times it’s served the same content to the same sort of people.

Automation is the process of applying technology to some repeatable task or process so that the task or process can be accomplished with predictable repeatability, lower total cost of operation, increase safety, and provide better efficiency.

This is what we demand of most of our technology, and technology has delivered that value. In fact, technology continues to deliver increasingly greater value to enterprises and individuals, squeezing more efficiency and capabilities and increasing productivity on a daily basis. So, automation is good. There’s nothing bad about it.

What is Intelligence?

However, we demand more from intelligence than simply automation. From the beginnings of what researchers have been attempting to do with Ai, we’ve been striving for systems that can understand and comprehend their surroundings, learn from their experiences, make judgements and decisions that are based on rational thinking, handle new situations and apply their learning from previous experience, and perhaps even address bigger questions of self-awareness, consciousness, and more. These are not easy problems Ai researchers are trying to solve, and are fundamental questions of what does it mean to perceive, understand, rationalize, and be aware.

The talk about the difference between narrow applications of Ai to single problem areas versus the strong / broad application of Ai which attempts to create a generally intelligent system. Clearly we’re nowhere near the goal of an AGI (Artificial General Intelligence), and so it would follow that all of the innovations in the market currently are narrow (some say weak, but we don’t) applications of Ai. Yet, that doesn’t mean that these narrow applications are any less Ai solutions. Image recognition, applications that reason with self-learning, and more are building the techniques we need to achieve general intelligence by creating solutions that apply theories of intelligence and learning to solve those problems. While they may be narrow, they are still intelligent and thus qualify as Ai solutions. But no one would argue that Facial Recognition is automation. Because intelligence is not automation.

Ask More Questions

Adding OCR or voice recognition of NLP (natural language processing) doesn’t make a system intelligent. Sure, the NLP library might be intelligent, but the system or solution could be just as dumb as ever, with a nice shiny new NLP system on it. Our old-hat integration or automation solution is still just an old-hat automation integration solution even if it has OCR (Optical Character Recognition). I could put an Alexa interface on top of my steam engine, but that doesn’t mean I can call my steam engine Ai-enabled. It’s still just expanding pistons and pushing rods. It’s automation, not intelligence. Talking to your car doesn’t mean your car is intelligent or Ai-enabled. Unless it’s autonomous, you’re still the one pressing the pedals and moving the steering wheel. You may be fooling some of your customers, your investors, your partners, and your employees, but you’re not fooling us.

Yet it’s worse than that. Vendors who push automation solutions as intelligence are hurting the industry. If customers are lead to believe that the automation solutions are what they can expect out of Ai systems are these weak automation systems that use the narrowest definition of Ai to call their systems intelligent, then the industry is heading for a rapid roadblock. If we want the desired outcome of Ai that we’re hoping for and expecting, in which we’re solving increasingly harder problems that have not before been able to be done with systems that can learn and adapt, then we have to ask more of the systems that call themselves artificially intelligent. When vendors say their products have Ai capabilities, don’t take them at face value. Ask how their offering is going to learn and adapt and perceive the environment. More than half of the time the response will be that they’re just doing automation with the thinnest possible veneer of Ai. The ones that are truly building Ai-enabled products in the way that will help the industry mature are the ones that are worth following. Let’s leave the hype train behind at the station. Where we’re going, there aren’t any tracks. (Apologies to Back to the Future for the misquote).

You Rock!

Avatar
Rihad Variawa
Data Scientist

I am the Sr. Data Scientist at Malastare AI and head of global Fintech Research, responsible for overall vision and strategy, investment priorities and offering development. Working in the financial services industry, helping clients adopt new technologies that can transform the way they transact and engage with their customers. I am passionate about data science, super inquisitive and challenge seeker; looking at everything through a lens of numbers and problem-solver at the core. From understanding a business problem to collecting and visualizing data, until the stage of prototyping, fine-tuning and deploying models to real-world applications, I find the fulfillment of tackling challenges to solve complex problems using data.

Next
Previous
comments powered by Disqus