Teach with expert insights on AI, curated by your trusty Teacher’s AIde

Welcome back to Teacher's AIed, where we don’t just sense that students are talking behind your back (remember, all teachers have eyes on the back of their heads), but also perceive that they are talking about you.

Are you smarter than a 5th grader?

Fill in the blank: Black and White Images are encoded as 2D arrays of pixels, where each pixel is a number indicating the _____ of that piece of the image.

What about this one:

What is the difference between perception and sensing?

According to AI4K12’s Big Idea #1, these are both questions that an AI Literate 5th grader should know the answers to. (Keep on reading to find the answers for yourself!)

In an age dominated by artificial intelligence (AI), it's not just students who need to grasp the fundamental concepts of how computers perceive the world. Educators, and indeed anyone living in this AI era, can benefit from understanding these principles.

This is the second installment in my series on AI literacy and its implications in K-12 Education.

Did you miss the previous post? Here’s the link:

• What Is AI Literacy? Part 1

In today’s post, I explore AI4K12’s first Big Idea - Perception. While some concepts may seem simple at first glance, they are crucial for comprehending the mechanics of AI.

What is AI perception? AI perception is the process by which computers and robots interpret sensory information from their environment. This involves using sensors to collect data and then processing it to make sense of the world.

It's akin to human perception, but instead of eyes, ears, and skin, machines use cameras, microphones, and other sensors.

Understanding AI perception is essential because it forms the basis of how computers and robots interact with the world. It's a cornerstone of computational thinking skills and a vital part of AI Literacy. As we delve into the three main concepts of AI perception—sensing, processing, and domain knowledge—we'll uncover the value of this type of thinking and knowledge.

1. Sensing

The first concept of AI4K12’s Big Idea #1 is Sensing: how do living things sense, how do computers sense, and how do computers store that information? Sensing is the first step in AI perception, where machines use sensors to gather data from their environment.

For a moment, let’s consider our bodies as if they were machines. We have all kinds of sensors built-in. I can still remember the little song from Kindergarten: “Sight, sound, smell, taste, and touch. Our five senses tell us sooooo much!”

This is intuitive and obvious (or to anyone past elementary school at the latest), but boiling down our anatomy proves an important point for AI Literacy.

If AI technologies approximate human competencies, then let’s understand what humans are capable of sensing and see how that could be recreated through computers.

Computer sensing includes visual information from cameras, audio from microphones, and even depth perception and motion detection from specialized sensors.

Consider self-driving cars, which rely heavily on sensing. They use a combination of cameras, radar, and lidar to "see" the world around them. This includes not just recognizing objects but also understanding their distance, speed, and trajectory. This level of sensing is far beyond simple vision; it's about interpreting complex sensory information in real-time.

This category also includes how sensory information is stored. Ready for the answer from the introduction? Black and White Images are encoded as 2D arrays of pixels, where each pixel is a number indicating the brightness of that piece of the image.

2. Processing

The second concept is Processing: how do computers transform sensory information into meaningful data?

This is where the real magic happens.

Once sensory data is gathered, the next step is processing this information to extract meaning.

This involves a hierarchy of increasingly abstract features and higher-level knowledge.

Let's take language as an example. As a human, I can't help but read and comprehend the words in front of me. My brain is hardwired to perceive language, not just sense it. Long gone are the days when I could appreciate script as just fancy squiggles (of course, unless I’m reading a language I do not know.)

For now and forever, I’ll be recognizing letters and understanding sentences. That’s processing.

In contrast, computers must be programmed to derive meaning from the squiggles. This process includes everything from extracting features (e.g., the lowercase letters “i” and “l” differ by a dot) to processing those letters as parts of larger words and then identifying the meaning of those words in the right context.

3. Domain Knowledge

The third concept is Domain Knowledge: how do computers use specialized knowledge to interpret sensory information? Domain knowledge is the specialized knowledge that AI systems use to interpret sensory information. It's what allows a computer to not just detect objects but understand their significance in a given context.

Going back to our self-driving car example, it's crucial that the car doesn't just see a yellow rectangle but understands that it's a school bus, which has specific implications for driving behavior. The car needs domain knowledge to recognize that regardless of its color or size, a school bus has certain characteristics and rules associated with it.

Slight Tangent: This is an interesting moment to bring in the third component of AI Literacy - moral and ethical implications. When considering self-driving cars, should they be programmed to treat school buses (or other high-occupancy vehicles) differently than other cars?

For example, if your car is driving in the middle lane and has to evade an upcoming road hazard. Your car checks its blindspots and sees a school bus filled with kids on the right and a luxury car with a single passenger on the left. What should your car do? Swerve into the bus? The Bugatti? Ram right into the hazard? Domain knowledge is how we instruct computers to handle ambiguity.

As we've explored, AI perception is a multi-layered process that involves sensing, processing, and domain knowledge. Understanding these concepts is essential for anyone living in an AI-driven world. By delving into the mechanics of how computers perceive their environment, we can better appreciate the complexities of AI and its impact on our lives.

Stay tuned for more in this series as we continue to unravel the intricacies of AI Literacy and its importance in education and beyond!

Class dismissed!

Lewis Poche & Kourtney Bradshaw-Clay