Given its recent popularity, you might think that AI or Artificial Intelligence is a relatively new field that emerged just a couple years ago. But what if I said that AI’s origins go all the way back to the 1950s? Alan Turing proposed the question, can machines think back in 1950? And John McCarthy defined artificial intelligence as the science and engineering of making intelligent machines in 1955. Obviously, a lot has happened since then, and especially since the release of ChatGPT in 2022 November.
The hype and advancements around AI have been huge. So, in today’s blog, I’m gonna explain AI, its related sub-areas, and what all of this AI revolution means to companies and their workforce in the most relevant and effective way to you. Whether you have no clue what artificial intelligence or machine learning or deep learning are, or you have quite a bit of experience and knowledge working in this field, make sure to stick around as I’m gonna bust some jargons and focus on what roles and skills you would need to acquire to be an integral part of AI-driven organizations.
So, AI first started out as a new branch of knowledge within computer science, a science focused on the automation of information processing by computers. And even if you’re not familiar with computer science, I’m sure you’ve heard of sub-areas within it such as programming languages or operating systems. Now, even though the massive advancements recently in the field of AI, there are still some pretty big limitations which can be easily explained and understood through comparing AI Artificial Intelligence to AGI Artificial General Intelligence.
The ultimate goal of AGI would be to build machines that mimic human-like intelligence, such as thinking, reasoning, or decision-making, to solve multiple complex problems. Now, it’s safe to say that we’re not there just yet. AI systems are good at solving one or maybe even a couple specific tasks, but Artificial General Intelligence does not exist yet. Think of voice assistance, face recognition, personalized product recommendation. These are all AI applications in our average, normal daily lives that many of us come across.
These AI applications more or less already work quite well. But now think of self-driving cars or generative AI with LLMs or Large Language Models, like ChatGPT. These are, I’d say, halfway toward real Artificial Intelligence.
There are certainly aspects that need to be improved upon. AI has various sub-domains, so I’ll only cover the most popular ones briefly. I’ll start with machine learning because it’s probably the one with the most hype around it at the moment.
At its core, machine learning, or ML, gives us the ability to learn from existing data, identify patterns, and make predictions or other inferences. Now, if you don’t know what any of this jargon means, don’t worry, as I’ll be simplifying all of this more technical content later on in the video. But for now, whilst I’m on the topic of machine learning, let me quickly draw your attention to the HarvardX Data Science Machine Learning course from edX.
And if you’ve never heard about edX before, it’s an affordable online learning platform with over 4,000 courses, flexible learning structures, so that you can find relevant programs and expert partners to support your every career moment, at any budget. edX was founded by MIT and Harvard professors and partners with top Ivy League universities to give you educational experiences to drive real professional progress. Choose the credentials that align with your career goals, and showcase your expertise to employers and peers, enhancing your professional profile.
So, by taking the HarvardX Data Science Machine Learning course, you’ll learn about popular machine learning algorithms, how to perform cross-validations, and how to build a recommendation system, amongst many other things, of course. This course is actually one of the 9 courses that the HarvardX Data Science Professional Certificate includes, a certificate that dives deeper into data science and data analysis skills and techniques.
So, within machine learning, deep learning is really popular nowadays. It focuses on training neural network models to solve the most challenging AI problems.
Now, robotics is another subdomain within artificial intelligence, and this one doesn’t need too much of an explanation, I believe. It’s focused on building AI-powered machines to do stuff like moving packages in a warehouse. Computer vision, another subdomain, investigates how AI systems visually interpret objects and Natural Language Processing, or NLP, which studies the ability to analyze, understand, and communicate information in a human-like way.
Just think about a smart warehouse system and how it works. The robots navigate the facilities, visually detect whatever objects they need, and communicate with each other to optimize all of the warehouse operations. It’s a combination of robotics, computer vision, NLP, and probably a lot more other subdomains.
Or think about the medical industry. Computer vision combined with deep learning can help identify various health conditions from images and videos like x-rays. Or just think about something seemingly super simple like smart home assistants such as Alexa.
It uses NLP and deep learning to communicate with us, give us answers, and recommendations. Now, on a little bit more technical note, let’s see what AI can and cannot do as of today. It certainly can make predictions and inferences and recognize patterns.
And some quick jargon busting here. Predictions mean forecasting something that is yet to happen, like weather forecasting for example. Inferences refer to determining a target output based on a set of inputs.
So, in plain English, think of every time when Netflix recommends a movie to you based on your watch history and preferences, that’s making inferences. Pattern recognition can include predictions and inferences, but also other tasks such as clustering. Think of customer segmentation to discover groups of data with similar characteristics.
Data generation or generative AI, which is as its name says, generates new data based on existing data patterns. Anomaly detection, which is commonly used in the financial services industry that I work in to detect fraudulent transactions, and is something that should definitely be used more by social media companies like Meta, for example. The amount of times I try to sell something on Facebook marketplace only to get all the scammers in the world tell me that their cousins, friends, wives, husbands, grandma, would come to collect the item, but they can only pay using PayPal, drives me insane.
Alright, that’s my rant over. Let’s move on to optimization and automation, things that AI can also do for us. Optimization simply means finding the best and most efficient possible solutions given specific constraints.
Think of finding the optimal delivery route in the supply chain and logistics domain. Dynamic pricing strategy for travel bookings on Airbnb to maximize revenue. Automation refers to following some rules to do some tasks repeatedly and autonomously.
Now, automation in itself is not AI per se, but AI can improve automation by improving the efficiency of the processes. Think of the robots, again, who manage parcels in warehouses. Or think of the process of screening a large amount of job applications.
Speaking of job applications, actually, have you checked out my most up-to-date resume templates? They can really help with your job applications. You can have all the skills in the world, but if you cannot even get through the first round when you’re applying to jobs, you won’t have the opportunity to showcase your skills and knowledge. So make sure to check out the resume templates.
The link is in the description below. Now, it’s clear that AI can already do a lot of things and do them well, but there are still some significant limitations to it. Conversational AI solutions, for example, are great at answering specific queries like, Alexa, how many grams in one ounce? One ounce is about 28.3 grams. But cannot yet mimic critical aspects of social interactions like humor, empathy, sarcasm, or emotional intelligence. Bias is another common issue which occurs when the AI systems make predictions that favor certain groups over others based on gender, age, ethnicity, or any other characteristics. This is obviously not ideal.
Also, AI systems are only as good as the data they’re trained on. Data is the fuel that feeds AI, and it’s also its biggest limitation probably. Without enough high-quality data to learn from, how effective are these AI solutions? Given all this AI talk, I think it’s time to put AI into an organizational, business, and industry context.
Companies will for sure try and establish some sort of an AI culture going forward to get a competitive edge over each other. AI can produce operational costs, automate boring, labor-intensive tasks, or improve customer experience. The benefits are endless.
And as these companies are trying to build their own AI culture, they’ll certainly need people who are AI literate at the forefront of their organizations to lead the AI revolution. Currently, the most popular type of AI infrastructure would be the cloud-based platforms like Amazon Web Services or AWS, Microsoft Azure or Azure, English is my third language so let me know in the comments below how you would pronounce it, or Google Cloud Platform or GCP. So, learning how to use these platforms and the tools and services that come with these platforms can certainly help you in the long run.
I actually included an end-to-end cloud project in my Ultimate Portfolio Template that I launched recently where I built an ETL pipeline in AWS using infrastructure as code. I’ll put the link in the description below, make sure to check it out if you’re interested. And as I said, organizations who want to become AI driven will need to hire the right people to drive the implementation of AI initiatives, people who can contribute to the successful adoption of AI technologies.
So, I’m gonna highlight a couple roles that are already in demand and will probably remain in demand for a while. Companies will definitely need AI architects to oversee the AI solution architecture. Think of people who have the expertise to make decisions on selecting the right AI tools to use.
Data scientists will also play a key role as you’ll need people who can prepare and analyze complex datasets, train and evaluate machine learning models and interpret model outputs. You’ll also need machine learning engineers who’d focus on deploying the implemented models into production and manage the AI infrastructure. Data engineers cannot be left out either of course.
They would specialize in building robust end-to-end data pipelines for the AI systems. And let’s not forget about the AI project managers and the AI leads who would be responsible for team and project leadership. Who knows, there may even be specialist ethics and AI domain expert roles to help companies responsibly implement AI systems and safeguard ethical considerations.
Now, I’m sure you noticed that I didn’t mention data analysts here, but that does not mean that data analysts would go extinct. Data analysts would probably become the end users of the AI systems and tools that these AI teams would establish. Thus, they would need to upscale as well to be able to use the latest AI technologies to extract meaningful insights out of data and answer business critical questions.
It’s clear to me that AI is disrupting and will continue to disrupt our society. It is reshaping the way we think, the way we do certain things, the way we solve problems. Artificial intelligence is evolving fast, probably faster than us humans can keep up with.
So I’m very curious to see what the future has to offer in this space. For now, I hope that by busting some jargons around AI, its related subdomains, its use cases, and introducing some AI skills and roles, you were able to get a better understanding of what artificial intelligence actually is, and maybe even think about the skills you need to pick up going forward to make yourself AI ready. If you enjoy blog like this, make sure to check out some of my other blogs.
Thanks so much for reading, and I shall see you in the next one.