What is Artificial Intelligence?

tl;dr: Artificial Intelligence, AI, has been around for fifty years, give or take, though we are just now beginning to hear the term dropped in every sales pitch possible. How do marketers stay informed of the advances in technology? This article helps break down AI into easy to understand talking points.

If you attend any MarTech conference, you’ll see rows of companies lining the hallways touting their AI capabilities and solutions, and it may be hard for marketers to make sure they’re not being sold snake oil.

Below are some helpful tips to understand what Artificial Intelligence is, how the technologies within AI differ from each other, along with some handy shortcuts to remember how it all fits together. This information will help you make sense of this technology and ensure you ask your vendors the right questions.

First, some history.

For the last decade or so, one of the major milestones in business has been the collection and use of ‘big data’. Massive amounts of information around spending, user experience, and digital behavior has been housed and used to evolve marketing to be more personalized, among other things.  

The next logical step after the big data revolution was the use of data for Artificial Intelligence and machine learning, both of which require large volumes of data.

In 1925, an inventor by the name of Francis Houdina once demonstrated a radio-controlled car. While the idea and opportunity for driverless cars existed in 1925, the technology and power did not, so the idea didn’t scale. Fast-forward to present day where the technology now exists allowing self-driving cars to flourish.

Americas Independent Electric Light and Power Companies placed an ad in LIFE magazine in 1956 which to this day gives one of the most detailed and estheticized depictions of autonomous driving. Photo credit link.

Artificial Intelligence and big data work in a similar way. AI and ML have been around for the past 50-60 years, becoming much more popular and relevant as of late. Why? Much like the car example above, the technology was there but it could never scale due to a lack of big data. The main reason why AI can be scaled today is from the work many industries have done regarding big data.

Easy Ways to Understand AI and ML:

The main crux of both AI and ML resides in their automated learning capabilities. The AI-solution that I developed at Autodesk accelerated conversions for paid search digital campaigns and we had to spend time training the machine. Training the program was key and made it unlike any other program you may have come across in marketing. Once the basic rules were set—depending on the type of technology being used—the machine took over and began “teaching itself”.

While AI has always done well with structured problems, it has struggled with reasoning. Reasoning is required for a range of cognitive tasks including using basic common sense, dealing with changing situations, simple planning, and making complex decisions in a profession. “Humans know that if you put an object on the table, it’s likely to stay on the table unless the table’s tilted. But nobody writes that in a book—it’s something implicit. Systems don’t have this common-sense capability,” explains Aya Soffer, IBM Director of AI and Cognitive Analytics Research, in a recent article. That issue is rapidly changing as the technology is evolving and more branches are being developed and introduced.

Artificial Intelligence can broadly be broken down into five major categories: rules based, machine learning, natural language, computer vision and artificial reality.

Below is a quick primer to help you understand the core advantages of these categories and how they can be applied to your world:

1.      Rules-based

What it is: A series of “if-then” statements telling the machine if it encounters x, then respond with y. Little is left to the machine in terms of variability, so if it encounters something not written in its rule set, it may error out.

Think: Turbo Tax and how the software uses rules (laws) and your personal taxes to figure out your tax liability.

2.      Machine learning

What it is: The use of algorithms and mathematical models that the computer uses to process a task and progressively get better at it. Deep learning lives here as a sub-category and there are many great things happening in the fields of medicine, safety, and more. Thanks to this technology machines can help us avoid dangerous circumstances, like drowsiness detection for drivers, or understand triggers of major problems, such as the onset of Alzheimer’s.

Think:  Current programmatic display whereby you ask the machine to find the best performing traffic to show your display ad to. If you were to shut down the programmatic display, you would essentially have to start over with the learnings. To get the machine to learn and find the best converting impressions, you must let it run for long periods of time, without disruption and with more data to allow it to learn and improve.

Rules-based and machine learning create the foundation for understanding the other categories of Artificial Intelligence.

3.      Natural Language:

What it is: This category includes processing, querying, and understanding data for interaction. Big data plays a major role in teaching the machine the variety of potential responses or solutions that it can give.

Think: Alexa, or the Autodesk AI-chat bot I created that accelerates conversions.

4.      Computer Vision:

What it is: Machine interpretation of images and videos, as well as both static and dynamic images. This is primarily used for object recognition and segmentation.

Think: Facebook facial recognition to automatically make tag photo suggestions.

5.      Artificial reality:

What it is: Augmented, mixed or virtual reality focused on the immersive experience. Some of the cool devices we play around with these days fall into this category.

Think: Nintendo Pokémon Go App.

As time goes on, we will see more Artificial Intelligence applied to our personal and professional lives, automating many areas and enriching them as well.