I2C SPI USB CAN eSPI Cable Testing View All Videos Quick Start Guides Software Downloads App Notes White Papers User Manuals Knowledge Base Sales Support About Us
Products Blog Sales Support Contact Search
Machine Learning and Artificial Intelligence (AI) – Examples, Pros and Cons
Brendan Murphy

Artificial Intelligence is a term that most people in the modern world are familiar with. Probably more commonly referred to as AL, artificial intelligence is a hotly debated topic. Do we want AI in our everyday lives, or do we not? We have heard many great things about how artificial intelligence can enhance and bring tremendous value to our daily lives, but we have also heard the flip side of that comment; the Orwellian, apocalyptic narrative. What if artificial intelligence decides that it would be best if humans ceased to exist; now, that wouldn’t be good, would it? But what is machine learning and AI precisely, and what are the pros and cons of the technology?

Artificial Intelligence, as the name directly implies, is intelligence that can be programmed or construed artificially, usually in a computer system. AI can also be thought of as a computer with human-made thinking power that enables a machine to simulate human intelligence and problem-solving skills. When given a problem or question, the device can respond with relevant information to help solve or answer the question. Artificial intelligence is programmed by “feeding” a computer system a large set of data and training it to recognize patterns in the data to respond to questions about the data accurately and intelligently. The data “feeding” or training process is done through a mechanic called machine learning.

circuits in shape of a brain Photo by Steve Johnson on Unsplash

Machine Learning

Machine learning is an application of AI that allows a system to train, learn, and improve based on experience rather than being instructed to do so. To do so, a group of engineers will establish the purpose of a specific AI application and build out clear parameters for it. They will specify what the AI will do and will teach it to learn everything there is to know about the subject by feeding it information. The information can come in data sets, photos, videos, and just about anything with a digital footprint. The information is then given to the program to analyze, review, and learn from, usually in the form of pattern recognition and model building. The machine is taught to look for specific aspects of the data set and then provide inferences on what it has gathered. Over time, the machine will use this data and a set of algorithms to imitate the way that humans learn, gradually improving its accuracy.

Examples of AI

Now that we have discussed the symbiotic relationship between machine learning and AI, what are some real-world examples of these technologies at work?

Voice Assistance

One of the most prominent examples of AI is modern-day voice assistance, such as Google Assistant, Amazon Alexa, and Apples Siri. These voice assistances will respond to questions when asked. Perhaps you have asked your phone something like “What’s the weather today?”, “Take me to the store.”, or “Call Mom.”, and sure enough, almost instantaneously, your phone will accurately perform all of these requests. These AI systems have been fed millions of data points and algorithms to learn how to best deliver some of the most common requests cell phone users have.

round grey speaker/voice assistant on brown shelf Image by John Tekeridis via Pexels

Search Engines

Every day, most of us use the internet in some form or another. Whether it be checking the news, looking for new dinner recipes, or planning your next vacation, a search engine probably plays a significant part in your internet exploring. The most notable and popular search engine is Google. The Google search engine is a prime example of an AI system that works to find personalized and specific information that it thinks you would like to see.

The internet is a powerful search engine composed of millions of recourses that allows us to find the exact information we are looking for. Google Search learns about your habits, preferences, locations, and much more in order to give you information that it thinks you will like best. The machine learning aspect of Google AI comprises reading and analyzing your search history, the amount of time spent on webpages, geographical location points, and much more. Once assessed, the AI has an excellent understanding of who you are and how to make your experience with the search engine enjoyable and easy to use.

Online Chat Bots

Perhaps you have done some online shopping and noticed that during your time on the site, you were able to seemingly chat with someone about questions you had about a specific promotional deal or shipping questions. Chances are you were talking to an AI system created to answer product-related questions. These chatbots are programmed to analyze common questions people have and then respond appropriately to them.

Pros and Cons of AI

The use and rapid adoption of AI has introduced a variety of costs and benefits to society. Generally, AI is considered a net benefit, but some worry about the potential of AI getting too intelligent. Let’s discuss some of the pros and cons of the technology.

Pros of AI

Time Saver

One of the most significant advantages of artificial intelligence is its ability to automate mundane and rather time-consuming tasks. Automation means less time spent working on “low-level” problems and more time working on other, more important tasks.

Less Human Error

Self-driving has become a buzzword in the automobile world in recent years. The reason companies are looking into self-driving is simple; robots can be better at driving cars than humans. Although full-fledged, self-driving cars are not a reality just yet, there will be a time in the future when they will be the mainstream. There are an estimated 6.5 million car accidents annually, and they say 98% of those accidents are due to human error, not breaking at the right time, looking down at a cell phone, etc. Taking humans out of the driving equation and letting a car drive itself could reduce car accidents to less than 130,000 a year.  This stark difference in numbers is exactly why self-driving is being invested in so heavily.

Safer Work Environments for Humans

AI can be used to program robots to do jobs that we deem as dangerous or hazardous for humans. We can use robots to do things like mine coal, defuse bombs, put out fires, etc. Robots can be used to do the most dangerous jobs in our societies, freeing up people to work in safer work environments.

Cons of AI


One of the most significant concerns around AI is the fear that it will diminish availability of jobs in society. If AI can be used to automate certain job functions, it could also be to the detriment of the worker whose job is being automated away. What do all the taxi and rideshare drivers do if cars can drive themselves? What about semi-truck drivers? How can measure up against a robot that can perform the job 24/7? Although the automation of jobs like these can be seen as innovative, there could be disadvantages for the workforce itself.

Rogue AI

Terminator, a sci-fi action film created in the 80s, is a prime example of what some people fear AI can become: a killing machine whose mission is to take over society and ensure the human race ceases to exist. If we create a device that can think by itself, knows how to build things, and knows what’s best for its survival, what happens if that machine starts to think humans are a problem that needs to be fixed or, more starkly, eradicated? When a machine can learn and make decisions for itself, who is to say that it will make beneficial decisions for humans?

Tools Used to Create and Debug AI Systems

Let’s take a quick look at some embedded systems tools commonly used in AI development applications.

Host Adapters

Host adapters like the Promira Serial Platform and Aardvark I2C/SPI Host Adapter are very popular tools used to interface with the I2C and SPI protocols. At the heart of every AI architecture is a base protocol layer of code that builds the application's parameters. The Aardvark adapter and Promira platform are two tools that developers keep returning to because of their ease of use and functionality. These tools allow users to program microchips on a circuit board and enable users to interface with the protocols in master and slave modes.

Promira Serial Platform front


Protocol Analyzers

Protocol Analyzers like the Beagle I2C/SPI Protocol Analyzer and Beagle USB 5000 v2 SuperSpeed Protocol Analyzer allow developers to sniff a protocol bus in true, real time. The ability to monitor I2C, SPI, and USB bus traffic in real time helps developers easily find and eliminate bugs in their AI applications.

Beagle USB 5000 v2 SuperSpeed Protocol Analyzer


After learning that machine learning is a process within the realm of artificial intelligence and exploring some of the ways we see these technologies used in the world around us, we can confidently say AI is a technology of the future. Although concerns arise due to the rapid adoption of artificial intelligence, there are many benefits as well. If AI is pursued and implemented in ways that minimize the risks associated with the technology, we can see ourselves living in a much more efficient society.