AI Forward: 
Your Future, Faster

Artificial Intelligence is moving forward, but your help is needed.

Hello world! Welcome to AI Forward.

Artificial intelligence (AI) is rushing forward, and we want you there with it.

By 2028, Grand View Research projects the global AI market to expand by over 40% — with revenues forecasted to top $1.8 trillion worldwide. But Grand View notes that artificial intelligence is already making an impact in crucial industries: business, automotive, finance, healthcare, manufacturing, retail, and logistics. You could expand the list to include cybersecurity, transportation, media, and countless more.

global AI marketplace size chart

All this growth means that demand for workers skilled in AI — and machine learning, a subdiscipline — is skyrocketing. According to Indeed, this demand has more than doubled from 2015 to 2018, and it looks like it will continue to increase.

chart of AI job growth since 2016

Building on data from the US Bureau of Labor Statistics, Georgetown University’s Center for Security and Emerging Technology (CSET) forecasts the US’s AI workforce to grow by 8% from 2019-2029, adding one million jobs to an existing AI workforce of nearly 15 million and doubling the growth rate expected for the country as a whole over the same time-frame. Notably, this growth is concentrated in technical roles, where growth is expected to even reach 13% in some cases.

Universities have taken notice. In 2021, Stanford University’s Artificial Intelligence Index reported that over the four years leading up to 2020, higher-ed institutions across the world saw bachelor’s instruction in AI double and master’s AI instruction grow by more than 40%. The numbers here are telling. While many go on to seek advanced degrees in artificial intelligence, CSET estimates that upwards of 65% of the current technical AI workforce in the US have at most a bachelor’s degree.

The takeaway? A variety of different kinds of education — from bootcamps, to online degrees, to traditional bachelor’s, master’s, and PhD programs, to executive education programs — can give someone the tools they need to drive AI forward.

With growth like this surpassing some of the biggest American industries, it’s not surprising that you’re here to learn more. Maybe you’re researching next steps after getting your high school diploma, GED, or associate’s degree.

Maybe you already have a bachelor’s or master’s in AI or a related field and are looking to get back into the classroom, but you don’t know where to start.

Maybe you already have a degree in the social sciences or the humanities and are trying to figure out how to transition into tech.

Or maybe you already have a great career, but are looking into how you can get smart on AI to drive your business forward. Regardless of your reason, we’re here to help.

We started AI Forward with a simple goal: to match driven individuals with stellar educational opportunities in artificial intelligence and machine learning. But we understand that it’s not always as simple as choosing a program, putting in an application, and getting accepted. AI is a broad and rapidly evolving field. For newcomers and veterans alike, there’s a lot to digest: inscrutable jargon, fledgling subfields, myriad career paths, different software, and programming languages — and, of course, the many new educational programs popping up every day.

With so many factors in play, how do you make sure you’re setting yourself up for success and choosing an educational path that will put you on track to meet your goals? That’s where we come in. Here at AI Forward, we’re committed to providing you with well-researched, current, accessible guidance that will help, whatever your needs may be. And you don’t just have to take our word for it. Throughout the site, you’ll find contributions from students and experts alike who will share experience, advice, and the latest developments from the field.

Of course, we have to pay for all of this, which means that we partner with certain educational providers and may receive compensation when you request information from them or engage with their website. In this, we strive to be discerning and transparent. We only partner with schools and educational providers we would go to ourselves, and we’ll tell you when we are suggesting a partner institution. When available, we’ll also provide you with information like graduation and outcome rates, costs and financing options, and student reviews so you can decide for yourself.

With that said, let’s get started. If you’re brand new to AI, we recommend reading on. We’ll give you an overview of what AI is (and what it’s not), a look into how AI and AI professionals are impacting industries today, and lay out some resources to help you break in. If you have previous experience in the field, feel free to skip the basics and instead check out our guest columns or our master’s and doctorate program recs.

What is AI? (And what isn't it?)

Most people first encounter artificial intelligence at the movies. From 2001: A Space Odyssey’s Hal to Star Wars’ C-3PO, from Arnold Schwarzenegger in the Terminator series to Scarlett Johannson in Her: on the big screen, technologies  have long been thinking, feeling, acting — and, sometimes, even rebelling — in uncannily human ways. 

But despite its prevalence in sci-fi, AI is no fiction. For more than seventy years, scientists have been researching how to build machines that think and act intelligently. Facial recognition, digital assistants, autonomous cars, and your Netflix recommendations are just the latest products of AI’s long history.

Even with today’s concrete examples of AI technologies, what exactly artificial intelligence is remains an open question. In Artificial Intelligence: A Modern Approach, the textbook most commonly used in AI classrooms, Stuart Russell and Peter Norvig suggest that a “standard model” of AI would be one “focused on the study and construction of agents that do the right thing,” machines or technologies that act rationally in a given situation. Still, they allow that while this “rational-agent approach” has played a prominent role in AI’s history, it’s by no means the only approach and will likely not be the best one for AI going forward.

Past divergence, Russell and Norvig argue, arose from differences in how intelligence was defined along the axes of “human–rational” and “thinking–doing.” Some researchers considered intelligence “in terms of fidelity to human performance.” These include British mathematician Alan Turing, whose “Turing test” held that a machine was intelligent when it could fool a human into thinking that it too was human in a series of written questions and responses. Other researchers focused not on emulating human activity on the outside, but instead on mimicking human thought processes themselves, and so tried to translate how humans think into computer code.

Still others sought to arrive at artificial intelligence by pairing the ability of computers to think rationally — that is, according to the laws of logic — with the theory of probability. “In principle,” Russell and Norvig explain, this pairing “allows the construction of a comprehensive model of rational thought, leading from raw perceptual information to an understanding of how the world works to predictions about the future.” Finally, there were those proponents of the rational-agent approach, who occupied themselves with creating machines that could act rationally: take input from the environment and output logical, goal-oriented behavior.

Thinking Humanly

“The exciting new effort to make comput- ers think . . . machines with minds, in the full and literal sense.” (Haugeland, 1985)

“[The automation of] activities that we associate with human thinking, activities such as decision-making, problem solving, learning . . .” (Bellman, 1978)

Thinking Rationally

“The study of mental faculties through the use of computational models.”

(Charniak and McDermott, 1985)

“The study of the computations that make it possible to perceive, reason, and act.” (Winston, 1992)

Acting Humanly

“The art of creating machines that per- form functions that require intelligence when performed by people.” (Kurzweil, 1990)

“The study of how to make computers do things at which, at the moment, people are better.” (Rich and Knight, 1991)

Acting Rationally

“Computational Intelligence is the study of the design of intelligent agents.” (Poole et al., 1998)

“AI...is concerned with intelligent behavior in artifacts.” (Nilsson, 1998)

Figure 1.1 Some definitions of artificial intelligence, organized into four categories.

In Artificial Intelligence: A Modern Approach, Stuart Russell and Peter Norvig orient kinds of artificial intelligence along the axes of “human–rational” and “thinking–doing."

While the rational-agent approach forms the basis of AI’s standard model,  as AI capabilities improve, the goals we have for the system are becoming less and less clearly defined. A model focused on building goal-oriented systems is thus less useful than it has been in the past. Instead, Russell and Norvig argue, tomorrow’s developments in artificial intelligence will have to reckon with the value alignment problem, the “problem of achieving agreement between our true preferences and the objective we put into the machine.” This makes sense when we consider some of the AI technologies we’re already encountering in our daily lives. 

In The Alignment Problem, Brian Christian points to myriad instances where AI technologies achieved their specified goals only to discriminate against part of the population, such as when images of Black individuals were mislabeled as “Gorillas” by Google Photos or when the second-most highly ranked word of a resume-screening application happened to be the man’s name “Jared.” Such mishaps are all the more reason why the AI field needs people from a wide variety of backgrounds. 

Broad representation of who we are as humans will help align our values with those of the technologies we are developing.

robot and human finger touching

Image credit: Brookings Institute

Weak vs. Strong AI

While different kinds of artificial intelligence can be defined, as Russell and Norvig do, along the axes of “human-rational” and “thinking-doing,” a further distinction is between so-called strong and weak AI, terms that refer not so much to the nature of the goals set for AI technologies as the scope of the desired abilities. According to IBM Cloud Education, Strong AI — also known as artificial general intelligence, or AG I— refers to AI systems with broad intelligence on par with, or even exceeding, human intelligence. Strong AI systems would be able to learn and successfully complete a wide variety of tasks, and potentially even develop self-consciousness. As of 2022, strong AI remains stuck in scientific theories and sci-fi books and movies: no examples exist, and, as IBM notes, some scientists even doubt that such technology is possible. Weak AI — also called narrow AI — refers instead to the kinds of task-based AI technologies quickly becoming ubiquitous: Google Translate, IBM’s Deep Blue chess computer, or Uber’s ride-hailing algorithms. 

Artificial intelligence is often also broken down into four subtypes of increasing complexity — reactive machines, limited memory, theory of mind, and self-aware — with the first two referring to existing “Weak AI” technologies while the last two refer to “Strong AI” capabilities that have so far been realized only in research journals and Hollywood films.

Reactive Machines

When people talk about “reactive machines,” they are talking about AI systems that have been programmed to select from a fixed (if very large) set of outputs when fed with an environmental input. As opposed to the more sophisticated subtypes of AI, reactive machines don’t learn, they only react. Examples of such systems include IBM’s Deep Blue chess computer, which famously beat grandmaster Garry Kasparov in a six-game match in 1997.

chess match ibm vs human

IBM’s Deep Blue battles chess grandmaster Garry Kasparov in 1997. Image credit: Adam Nadel/AP Images

Limited Memory

As opposed to reactive machines, limited memory AI systems can remember recent experience and use this experience in decision-making. A good example of a limited memory system is a self-driving car, which must be able to assess its surroundings not just at the moment a decision is made, but also in the near past to assess the speeds and trajectories of other cars (and pedestrians!) on the road.

Theory of Mind

Theory of mind AI systems are so-named because they have an understanding that they are interacting with humans that have “minds”: emotions, thoughts, and moods. Although none have been developed yet, if developed in the future these systems would not just passively serve humans or interact with them at the level of discrete tasks, but instead actively cater to the particular mental states of their users.

Self-Aware

Self-aware systems would move beyond mere emotional interaction with humans to themselves develop complex consciousness. Such systems would be able to learn autonomously, reflect on their being, and maybe even have their own emotions.

Machine Learning vs. AI

Machine learning has found such widespread application these days and sounds so similar to artificial intelligence that it’s easy to confuse the two. 

The distinction is quite simple: machine learning is a subfield of artificial intelligence that employs mathematical algorithms that allow computers to progressively improve their capabilities. Russell and Norvig state the steps quite simply: “a computer observes some data, builds a model based on the data, and uses the model as both a hypothesis about the world and a piece of software that can solve problems.”

Machine learning scientists and engineers generally set up a computer to learn by feeding an algorithm a “training set”: a vast trove of data that helps the computer develop its predictive model for future situations it will be faced with. If you’ve ever transcribed a word or identified stop lights or bicycles for a reCaptcha, you’ve contributed to a training set that will help digitize books and allow cars to drive themselves.

As with AI, machine learning comes with its own set of specialized terms. Chief among these are the (related) terms “neural network” and “deep learning.” Neural networks — in longhand also called “simulated neural networks” (SNNs) or “artificial neural networks” (ANNs) — are systems of algorithms inspired by the synapses and neurons of the human brain. As you’ll learn in the video below, they frequently employ many different layers of nodes that, as they are trained over time, learn to only allow information to pass through them if it will help arrive at a correct output. “Deep learning” simply refers to the branch of machine learning that employs these “neural nets,” with the “deep” referring to the many layers of nodes.

Learn more about the difference between AI vs Machine Learning vs Deep Learning or take a deeper step into the Common Machine Learning Terms.

Machine Learning vs. Data Science

Data science is a hot topic in conversations about machine learning. Together they are two of the fastest growing tech fields today, and there’s substantial overlap — but it’s important to understand the differences. According to IBM, data science is “a multidisciplinary approach to extracting actionable insights from the large and ever-increasing volumes of data collected and created by today’s organizations. [It] encompasses preparing data for analysis and processing, performing advanced data analysis, and presenting the results to reveal patterns and enable stakeholders to draw informed conclusions.” 

From this description, we can see that data science is both broader in its approach to data and more narrowly focused on business goals than machine learning. Data scientists deal with data at all stages of the data pipeline: from cleaning and reading it for analysis; to making hypotheses about what analysis will show; to analyzing it, often with the help of machine learning models; to visualizing their findings to present them persuasively to interested parties. For their part, machine learning specialists focus on the actual design and implementation of algorithmic models that will learn over time, provide ever more accurate predictions, and thus allow for more insightful data analysis and task automation.

Careers in data science and machine learning have considerable overlap in skillset: for both, you need to be well-versed in mathematics like statistics and linear algebra, as well as programming languages like R and Python. Machine learning specialists will likely need to have deeper competencies in these areas, while data scientists will require skills in experiment design and presentation.

Learn more about the difference between Data Science vs Machine Learning.

charmin-rollbot

Employ the go: Charmin’s Rollbot provides on-demand TP delivery.

Other Subfields of AI

Machine learning has gotten considerable attention lately — and given what the future holds, this attention won’t wane anytime soon — but it’s not the only subfield of AI making waves today: natural language processing (NLP), computer vision, and robotics are all seeing exciting developments every day. 

Natural Language Processing

Researchers and engineers working in the subfield of natural language processing (NLP) are concerned with giving computers the ability to understand, and even employ, written and verbal language at a near-human level. In Artificial Intelligence: A Modern Approach, Russell and Norvig identify three goals of NLP: computers are being taught to understand and use human language 1) to communicate with humans, 2) to learn from the vast amount of information encoded in human language (like the internet), and 3) to advance the understanding of language we’ve developed in disciplines like linguistics and cognitive science.

Current commercial applications of NLP include translation tools like Google Translate, email spam-filters, personal assistants like Apple’s Siri or Amazon’s Alexa, and accessibility tools like Youtube’s automatic captioning. Another notable application is Generative Pre-Trained Transformer 3 (GPT-3), a predictive language model developed by OpenAI. The GPT-3 model can produce convincingly human-like text when fed short prompts, an ability K Allado-McDowell harnesses in the Pharmako-AI, a novel co-written with GPT-3.

Computer Vision

Computer Vision is an AI subfield with a predictable goal: to endow computers with the ability to perceive, assess, and act on visual stimuli at a level equal to, or even exceeding, human vision. This ability is essential if computers are to become even more active in the physical world in the coming years.

That said, computer vision is already impacting our lives every day. Perhaps the most ubiquitous example is Apple’s Face ID technology, a facial recognition system that precisely maps a user’s facial biometrics. Computer vision is essential for the development of autonomous vehicles, augmented reality video games like PokémonGo, as well as for a host of industries such as manufacturing and agriculture. Together with NLP, computer vision also plays a role in tasks such as text digitization.

Robotics

While robotics is a separate discipline unto itself, there is increasing overlap with AI as researchers look to introduce intelligent technologies into the physical world. To engage with the world around it, a robot relies on a combination of sensors and effectors. Sensors provide an input of environmental information, while effectors — robotic arms, legs, grippers, and the like —  allow a robot informed by these inputs to produce physical outputs and effect change on the world around it.

It’s well-known that robotics and AI have been driving increasing automation in a host of industries, perhaps most notably manufacturing. In recent years, however, the idea of AI-driven robots living with us in our homes has become closer to being a reality. Take, for example, the domestic robots highlighted in this report from the Consumer Technology Association’s 2020 Consumer Electronics Show (CES), which include Samsung’s rolling personal assistant, Ballie, Charmin’s toilet paper rescue-bot, Rollbot, and LOVOT, an emotional-support robot.

How is AI helping industries grow?

The examples above give a glimpse into some of the latest developments in artificial intelligence — but what’s the big picture? How is AI impacting our world in the twenty-first century? This section will give you an overview of how AI is driving essential industries forward the kinds of jobs that will let you get involved.

According to PricewaterhouseCoopers, AI technologies could be responsible for 14% growth in global GDP by 2030, the equivalent of $15.7 trillion. Driving this growth, PwC believes, will be productivity gains resulting from the automation of existing processes and the augmentation of the existing workforce, as well as increasing demand for the kinds of new and improved products that AI will make possible. This rapid growth means that industries like healthcare, finance, retail, and human resources will be ripe for AI-disruption. The report anticipates that “one of today’s start-ups or a business that hasn’t even been founded yet could be the market leader in ten years’ time.”

In this next section, you’ll find more information on how AI is already impacting essential industries — and how it might revolutionize them in the future. After that, we’ll look into what all this means for workers. Though automation necessarily means that some jobs will be phased out, you’ll learn about the kinds of new careers that will become increasingly sought-after as AI continues to transform industries.

How is artificial intelligence being used every day?

The PwC report provides a useful schema for thinking about how AI can positively impact productivity and product & service design for businesses by breaking this impact into four categories:

  • Assisted Intelligence: AI technologies that can help humans make decisions and perform tasks more efficiently

  • Augmented Intelligence: AI technologies that can help humans make better decisions and perform tasks better

  • Automation: AI technologies that can perform tasks usually performed by humans

  • Autonomous Intelligence: AI technologies that can learn on their own and determine new tasks and systems that will help meet business goals

Industries are leveraging these different forms of impact in a variety of ways to move forward. In finance, think tank The Organisation for Economic Co-Operation and Development (OECD) sees artificial intelligence, and machine learning in particular, positively impacting a variety of areas across the sector, whether “asset management, algorithmic trading, credit underwriting or blockchain-based finance.” On the nature of this impact, OECD is largely in agreement with PwC’s assessment of what will be behind AI-driven growth in the next decade: increased productivity and improved products and services. In finance, this will take the form of ML-optimized predictions and forecasting, automated trading, stronger risk-management, and bespoke product and service offerings. 

In healthcare, artificial intelligence will play a role in driving research and development of pharmaceuticals, providing improved diagnostics and testing, as well as boosting efficiency and security for day-to-day concerns like appointment scheduling and patient records. The Economist notes that these are welcome changes: the number of patients on earth is only growing, and the supply of doctors is struggling to meet this demand.

Accordingly, AI is projected to add $34.5 billion of value to the global healthcare market between 2020 and 2027, growth that could be accelerated or even exceeded if healthcare professionals themselves familiarize themselves with basic AI skills like coding and machine-learning model development.

Manufacturing is another industry where AI is already making a difference and will continue to do so in coming years. The effects of automation have already been felt in factories across the globe for some time. Increasingly, however, as PwC reports, artificial intelligence and machine learning will allow not just for the physical labor of manufacturing to be accomplished more quickly, efficiently, and safely, but for improved supply-chain management, more flexible production scheduling, and optimized product design.

What are the risks of AI impact in industries?

AI automation has revolutionized areas like manufacturing, customer service, and even copywriting. This unfortunately means that many workers have seen their positions made redundant. Time Magazine reports that AI automation led to the downsizing of 400,000 US manufacturing jobs from 1999 to 2007, a number which could swell by over 2 million by the end of 2025. COVID-19 has not helped: the US saw 40 million jobs eliminated at the height of the coronavirus pandemic. Time cites one group of scholars who believe that 42% of these positions will not return, with many replaced by increasingly capable AI.

Yet others are more bullish on the prospect of an increased number of humans working with AI in various capacities soon. PwC, for instance, optimistically rates the job-creating potential not just when AI is applied to industry, but in the continuing development of AI itself:

“The adoption of ‘no-human-in-the-loop’ technologies will mean that some posts will inevitably become redundant, but others will be created by the shifts in productivity and consumer demand emanating from AI, and through the value chain of AI itself. In addition to new types of workers who will focus on thinking creatively about how AI can be developed and applied, a new set of personnel will be required to build, maintain, operate, and regulate these emerging technologies.

For example, we will need the equivalent of air traffic controllers to control the autonomous vehicles on the road. Same day delivery and robotic packaging and warehousing are also resulting in more jobs for robots and for humans. All of this will facilitate the creation of new jobs that would not have existed in a world without AI.”

Who is this emerging AI workforce? What skills are needed to support the continuing development of AI technologies? What kinds of career trajectories do these professionals have?  And how can you put yourself on the right path? Read on to discover the answers to these questions and more.

What kinds of jobs are out there for people with AI skills?

Given the growth and impact projections for AI across industries, it’s no surprise that workers skilled in artificial intelligence and machine learning are already in high demand. As CSET explains, it’s difficult to quantify the labor shortage exactly, and estimates of the disparity between labor supply and demand vary widely. But if AI salaries are any indication, going into AI is not just a smart decision — it’s also a lucrative one. CSET cites a survey of Kaggle, a machine learning and data science community, that puts the median salary for US respondents at $110,000 per year, nearly triple the 2020 median US salary ($56,287). “Top AI talent in corporate labs,” CSET adds, “makes several times that amount, even in entry-level jobs.” That’s right: six-figure salaries aren’t only reserved for PhDs. 

Remember: the vast majority of the AI workforce has at most a bachelor’s degree.

A career in artificial intelligence or machine learning can take many different forms. Some choose to work as data analysts or data scientists, leveraging their familiarity with machine-learning modeling and programming to help businesses draw insights from their data, accurately measure past performance, and make predictions for the future. Others work as artificial intelligence engineers, designing, building, and perfecting systems and models for NLP, machine learning, computer vision, or other subsets of AI. 

In recent years, machine learning engineers have become especially sought-after as businesses have turned to neural nets and deep learning to optimize their operations. Another common path for AI professionals is to work as a software developer, software architect, or software engineer, creating and maintaining the software through which AI technologies are brought to businesses and consumers. 

Those taking a more academic route will often choose to become researchers at universities, think tanks, or corporate labs like those of IBM, Facebook, or Google. Researchers work on a wide variety of problems across such AI subsets as NLP, computer vision, and reinforcement learning.

While these different paths each require specialized expertise and experience — some more than others — they all require a set of core AI skills and competencies that are employed to greater and lesser degrees. Unsurprisingly, at the forefront are technical skills, including facility with software (e.g. TensorFlow, Apache Spark) and programming languages (e.g. Python, R, Java), a firm grasp of relevant branches of mathematics (in particular statistics and linear algebra), and proficiency in areas like signal processing and neural networks.

Also key are soft skills like critical thinking, written and verbal communication, and ability to work under pressure. Often, industry expertise is also necessary, particularly in roles where you will be expected to align AI systems and models with business goals.

How do we keep AI safe?

As AI capabilities continue to improve and the demand for AI technologies grows, so too will the need for AI policy research that ensures optimal outcomes for humans. In fact, 80,000 Hours, a non-profit careers-guidance outfit specializing in high-impact career paths, counts AI safety and policy among the most pressing problems facing our species. 

Chief among the concerns is the development of AGI, superhuman AI that would surpass human intelligence and potentially act in misalignment with our values. As 80,000 Hours lays out, there are a variety of ways to contribute to the development of smart AI policy and help improve AI safety. While familiarity with artificial intelligence and machine learning is a given, there are opportunities for individuals with a variety of backgrounds and interests — from government, to journalism, to nonprofits — to play a role in ensuring that AI will work for us in the coming years and beyond. If AI policy or safety interests you, we recommend that you check out 80,000 Hours’ guide.

How can you move forward in AI, and how can we help?

With the exploding demand for professionals with AI skills, it’s no surprise that universities and other educational providers increasingly seek to provide high-quality AI & ML instruction, both in existing computer science departments and through new, standalone artificial intelligence and machine learning programs.

Student growth in the field is noticeable from the undergraduate level on up. As mentioned above, Stanford’s AI Index found the number of courses offering undergrads instruction in developing practical AI and ML models to have grown by 102.9% over four academic years from 2016 to 2020.

Enrollment at these schools has also grown considerably in introductory courses in artificial intelligence and machine learning, increasing by 60% over the same time-frame. And Stanford notes that almost 30,000 students graduated with a computer science undergraduate degree in 2019 in North America alone, three-fold growth over 2010. While not all of these students graduated with AI- or ML-specializations, together with course-offering data this paints a picture of robust participation in artificial intelligence and machine learning at the undergraduate level.

chart of undergraduate graduates in North America from 2010-2019

While the growth in course offerings at the graduate level has been more modest — around 40% over the same four years — graduate programs have seen substantial growth in the number of tenure-track faculty who specialize in AI, with over 60 new professors joining the ranks of the 18 schools surveyed. 

The market is justifying that demand for qualified professors: as we mentioned above, computer science PhD students are increasingly specializing in artificial intelligence and machine learning. Indeed, as the number of CS PhD students tripled over the ten years leading up to 2020, students within this group working in artificial intelligence, machine learning, or robotics grew by over 10% and now make up over 20% of all new CS PhDs. Of course, these data miss the many graduate students working in the new AI and ML-specific master’s and PhD programs that are launching with every academic year.

Though it’s the most common path, traditional higher-ed is not the only route to a career working with the next generation of artificial intelligence. Increasingly, bootcamps and executive education programs (EEPs) are providing learners with the skills and expertise they need to break into the field or supercharge their businesses. Like traditional degree programs, bootcamps and EEPs offer rigorous curricula, stellar faculty, thought-provoking guest speakers, and ample opportunity for networking. At the same time, these programs generally focus more squarely on directly marketable skills and the various applications of AI to industry, with shorter durations and more flexibility due to being predominantly online. 

This is not to say that those seeking a bachelor’s, master’s, or even PhD are out of luck should they also desire fast time-to-degree and flexible hours. Encouraged in part by the revelations of the pandemic, more and more brick-and-mortar higher-ed institutions are offering students the option to complete their studies partially or wholly online. As would be expected, the US Department of Education’s National Center for Education Statistics (NCES) reports that colleges and universities experienced a 93% rise in enrollments for “distance education courses” — online learning — from the fall of 2019 to the same time in 2020. 

More surprising, however, is how higher-ed institutions plan to move forward after the pandemic. Data from the National Council for State Authorization Reciprocity Agreements (NC-SARA) cited in the same InsideHigherEd article suggest that of the 2,200 institutions surveyed, 59% will continue to provide opportunities for online learning after the pandemic. This is supported by the more bullish forecast by Wiley Education Services and EducationDynamics — a forecast based on data gathered before the pandemic hit the US — that “more than 70% of colleges and universities expect to launch one to four new online undergraduate programs over the next three years.”

This is great news for those seeking to complete a traditional AI degree while continuing to work, care for loved ones, or reduce housing and travel costs.

Encouraged in part by the revelations of the pandemic, more and more brick-and-mortar higher-ed institutions are offering students the option to complete their studies partially or wholly online.

At this point, you might already have a good idea of what kind of degree or certification you’re looking for, but others likely still have questions: 

  • What can I expect to learn?

  • Should I study artificial intelligence or machine learning within a computer science department, or should I seek a specialized AI or ML program? 

  • Do I need to get a full bachelor’s degree in AI, or will a bootcamp suffice for my goals? 

  • What can I do after a master’s or PhD degree?

  • Is there an EEP tailor-made for my industry? 

You might also have logistical questions about the application process, financial aid, or online study.

In the next section, we’ll address all of these questions and more, diving deeper into the AI and ML educational opportunities available today.

What educational paths are out there?

Undergraduate

Bachelor’s Degrees

Students looking to study AI as part of a Bachelor of Arts (BA) or Bachelor of Science (BS) course of study have options when it comes to picking a major. While most instruction in artificial intelligence and machine learning is still taking place through AI or ML specializations added onto traditional computer science majors, standalone AI and ML majors are becoming increasingly popular.

So which is right for you?

Luckily, the choice isn’t as crucial as you might think: the two share remarkably similar curricula. 

With some exceptions, bachelor’s degrees are designed for students who have aptitude in a particular subject, but little or no previous post-secondary educational experience. This means that regardless of whether you choose to major in AI or CS with an AI specialization you can expect to take foundational courses in math, statistics (e.g. Multivariable Calculus, Linear Algebra, Probability), and computer science (e.g. Computing Systems I&II, Principles of Functional Programming, Database Systems, Principles of Software Engineering), as well as AI-specific introductory courses and electives (e.g. Introduction to Artificial Intelligence, Deep Learning, Robotics).

Many programs also require a capstone project to be completed during your final year of study. Before that you can also expect to take a course in AI ethics and policy, a writing and communication course, as well as distributional requirements in the social sciences and humanities. Often, undergrads will also have the opportunity to assist their professors in cutting-edge research. And while a bachelor’s in artificial intelligence is an intense course of study, you might be able to complete it from your couch: increasingly, schools are allowing some or all of this coursework to be completed online. 

Graduating from one of these bachelor’s programs generally means you can enter the job market immediately, working as a junior artificial intelligence or machine learning engineer, data scientist, or software developer. Or you might be recruited for a corporate training program, where you would rotate through different roles to gain experience and a better understanding of what might interest you. Another option would be to enter graduate study right out of undergrad. 

Though there is significant demand in the labor market for AI professionals, there is still substantial competition for roles. Accordingly, it’s important to pay attention to student and alumni outcomes when researching schools. It’s also a good idea to choose a program that offers robust career services and active alumni networks, which are invaluable assets whether you are just starting your career or looking to move up.

Other important things to take note of when researching schools are cost and financial aid options. Tuition for public universities will generally be lower, especially if you are a resident and entitled to pay in-state tuition. Oftentimes, online programs will also have lower tuition. But regardless of a school’s tuition, there are many financial aid options available to you, from private scholarships and grants; to federal grants, scholarships, loans, and work-study programs; to private loans. In fact, over 80% of US undergrads receive some form of financial aid, so you’re in good company if you’re seeking help paying for school. The most important thing is to inform yourself about your options — and utilize private loans, which often come with exorbitant interest rates, only as a last resort. 

Of course, before paying for school, completing a course of study, and entering the AI workforce, you’ll have to apply to a school and be admitted. While the process is similar for most schools, each has its own particular requirements and deadlines, so as you are researching programs you’ll want to take note of these details. To save you time, we’ve assembled pertinent information about stellar programs on our Undergraduate Degree Application Details page.

When applying to a program, you can generally expect to be evaluated for admission based on all or a mix of the following:

  • your achievement in high school

  • your standardized testing scores (including for tests like TOEFL if you are not an English native speaker)

  • the interests, character, and ambition demonstrated by your extracurricular activities

  • recommendations written by your high school teachers and guidance counselors,

  • your written answers to essay prompt

  • An interview with an alumnus, alumna, or admissions officer 

Head over to our article on AI degree requirements for a deeper dive on this topic.

For AI programs, you will very likely need to show aptitude or readiness for the kinds of advanced math and statistics you will be required to learn, either through prior study in high school, AP tests, or through math boot camps, like AI Plus’ Foundations of Artificial Intelligence, or massive online open courses (MOOCs) like Imperial College London’s Mathematics for Machine Learning. Most programs will also require you to pay an application fee. But again, schools vary. Some schools are more focused on curating learning communities, and so might more heavily weigh essay responses that help them get to know you. Others might have temporarily or permanently lifted standardized testing requirements in light of COVID-19. Still others might have waived their fee requirements. When in doubt about what might be required, check!

Graduate

If you’re seeking a substantial course of study to help you advance in the world of AI and ML, and you already hold a bachelor’s in computer science, artificial intelligence, another STEM field, or even the social sciences and humanities, then a master’s and PhD program in computer science, artificial intelligence, or machine learning is a great option. Graduate study offers more advanced coursework, increased access to professors, and the opportunity to specialize in a subfield of AI or machine learning. For all PhD programs and many master’s programs, you will also be expected to undertake original research culminating in a dissertation, thesis, or capstone project — research that will often play an important role in recruiting. But should you opt for a master’s program or go directly into a PhD? We’ll lay out some of the differences so that you can make an informed decision. 

Master’s Degrees

Your experience in a master’s program in artificial intelligence or machine learning will largely depend on the knowledge you bring into it. If you are coming in with a bachelor’s degree from a field far removed from computer science, AI, or machine learning, you will likely need to complete what is often called a “conversion master’s” — or at least conversion courses — which will furnish you with the skills and expertise that would be expected of somebody with a bachelor’s degree in computer science, AI, or machine learning. In a conversion master’s, you will have some opportunity at the end of the program to begin more specialized study. If you already have fundamental knowledge in the field, you will be able to both deepen this knowledge and begin specializing from the get-go.

But why get a master’s in the first place?

If you already have experience in the field, the primary reason is practical: certain lucrative positions (senior AI or ML engineer roles, for example) in Big Tech and other leading companies are usually so competitive that you can really only get an interview if you have a master’s degree.

Another reason people enter a master’s program is that they hope to eventually start doctoral studies but don’t yet have sufficient credentials to be accepted to their choice of programs. Whatever the reason, it’s important to be realistic about the financial impact.

While a master’s might open up more lucrative salaries or further study, you should consider how the costs associated with going back to school and two or more years out of the workforce (and so two or more years foregoing promotions) might impact your lifetime earnings. It’s very possible that the higher earning potential that comes with an advanced degree in AI or ML will more than make up for any temporary financial setbacks, but you’ll be happier with your choice in the long run if you make it with eyes open to the financial realities.

Another option to consider is if an online or part-time program would allow you to keep working, and thus avoid the considerable financial impact of studying full-time in person. While places in such programs are competitive, you might also try to gain admission to one of a handful of partially- or fully-funded master’s programs in computer science that allow you to pursue artificial intelligence while receiving a tuition remission and/or earning a modest stipend for serving as a research or teaching assistant.

Interested in online master's programs? Head over to our deep-dives:

While much of the application process for artificial intelligence and machine learning master’s degrees are similar to the process for bachelor’s degrees, there are some key differences. In addition to an online application (and usually an application fee), you will generally be required to submit a personal statement detailing your experience, research interests, and reason for applying to a particular school. You will also need to submit up to three recommendations, usually from undergraduate professors or professional acquaintances, as well as a resume or CV, transcripts from all post-secondary study, and standardized test scores.

The scores most commonly accepted are from the Graduate Record Examinations (GRE), which comprise an analytical writing section, two sections each testing your verbal reasoning  and quantitative reasoning, and usually also two ungraded sections used for test calibration and research. As with the SATs, some schools may have temporarily or permanently waived the testing requirement. Lastly, though they are usually not required, it’s a good idea to inform schools of any publications of original research in leading journals or presentations at notable conferences: these will certainly support your application.

Doctoral Degrees

If you want to conduct research in academia, at a think tank, or in industry, you will need to earn a PhD in artificial intelligence or machine learning. The percentage of graduating CS PhDs who specialize in artificial intelligence and machine learning has grown dramatically over the last ten years: Stanford’s AI Index reports that such students made up over 20% of all CS PhD students in 2019, up 8.6% from 2010 and dwarfing all other CS specializations. At the same time, pursuing a PhD in artificial intelligence remains an exclusive club. With the rigorous requirements and lengthy time-to-degree (generally 4-5 years), it’s no surprise that the same survey counted less than 300 new PhD students specializing in AI or ML graduating from American computer science departments in  2019.

In the US, students showing promise are frequently admitted to PhD programs without a master’s degree, with the understanding that they will complete a master’s course of study (and the requisite comprehensive exams) before being allowed to begin doctoral research (a process usually known as “qualifying”). As with bachelor’s and master’s programs, if you want to specialize in a subfield of artificial intelligence or machine learning, you can do so within a traditional computer science department or in a PhD program that specializes in AI — the key is to choose a program that has faculty with similar interests to you that will be able to guide you as you progress in your research. When looking into programs, make sure to read individual professors’ work to see if one or several of them might be a good match. Oftentimes, professors will also happily answer your questions or even meet with you if asked over email. It’s also a good way to connect with them and start building a relationship. You can also check our PhD program profiles to get a better idea of where you can pursue your research interests.

While most students still pay for computer science and AI master’s degrees in the US, the majority of PhD programs are fully-funded, meaning that PhD students don’t have to pay tuition or fees for their PhD and the master’s degree(s) they earn along the way, while also receiving a stipend to study and conduct research. If you’re passionate about research, this might seem like a dream come true, but remember that this stipend will be far lower than the salary you would earn in industry.

Increasingly, however, researchers believe that a temporary deferral of a high salary is worth it: while the traditional path for a successful PhD candidate would have them entering academia as a junior professor, potentially after one or more stints as a postdoctoral researcher, increasingly PhD graduates are going into industry, where astronomical compensation awaits. This shift is in part due to the saturation of academic departments: as Stanford’s AI Index explains, the 48% increase in the share of PhDs going into industry between 2010 and 2019 isn’t due to fewer PhDs entering academia so much as “the large increase in PhD output [...] primarily being absorbed by the industry.” But with the exciting research coming out of places like Google’s DeepMind and Meta AI, no teaching requirements, plus the perks that come with working in big tech, you can imagine that many new PhDs are only too happy to enter industry.

Short-Courses: Bootcamps & Executive Education Programs

Maybe you’re looking to get smarter about artificial intelligence and machine learning, but you’re not looking to enter a lengthy (and costly) degree program. Luckily for you, shorter programs that can help almost any education-seeker meet their goals are becoming increasingly available. For those with existing technical skills in mathematics, programming, or software development looking to expand their skill-set, many universities and other education providers are now offering machine learning bootcamps and online ML certificate programs that will allow you to draw on this experience as you begin to develop technical expertise in AI and machine learning. While these courses usually won’t award college credit or formal degrees, they are often part-time and online, allowing you the flexibility to continue working as you upskill.

Options also exist for those without technical experience just looking to get a better handle on the AI landscape and how AI technologies can help their business. Whether you’re a C-suite executive looking to build out your company’s AI footprint, a senior manager looking to better lead your technical teams, or a mid-career professional looking to improve your career prospects for the future, executive education programs can help you understand the past, present, and future of artificial intelligence and machine learning in your industry, be it business and finance, healthcare, manufacturing, or any of a host of others. As with the more technical programs, these courses will likely be part-time and online, allowing professionals to study on their schedule without disrupting the day-to-day of their business.

While short courses generally cost significantly less than degree programs, there are still many kinds of financial assistance available for both technical and non-technical courses. Many programs offer need-based scholarships, flexible payment plans, or even group discounts. Increasingly, bootcamps are also allowing students to pay for their education only after they have gotten a job with a certain level of compensation. If you are currently employed, you might also check to see if your employer would cover some or all of your educational expenses: many companies offer workers tuition or education reimbursement as part of their employee benefits. As for degree programs, private loans should be your last resort to finance your learning.

How do you land a job after school?

Finding, getting accepted to, and completing an AI degree program or short course is a great way to start breaking into artificial intelligence, but it’s only half the battle: even before you’re out of school, you need to focus on landing a job. The high salaries and massive growth potential that make a career in artificial intelligence and machine learning appealing have also made the job market very competitive, even if new positions are opening every day. So what can you do to set yourself apart?

The first step is a resume that lists your relevant experience and achievements from coursework and any internships you completed during your studies. For most artificial intelligence and machine learning positions, you will also need to provide a portfolio that showcases AI and ML projects you have worked on. Portfolios usually take the form of a Github repository — an online folder that holds project files, their revision histories, and explanatory READMEs — or a blog-style personal website. But just where can you find projects to work on that will help you stand out from the crowd? Oftentimes, you’ll have the opportunity to work on projects that interest you in the later semesters of your degree, but you don’t need to wait! Hackathons, open-source projects, and coding competitions like those found on Kaggle are a great way to sharpen your skills and build up your portfolio.

It’s also a good idea to start building your network. You can start by holding short informational interviews with friends, family, and other graduates from your program and branch out from there. Networking like this is a great way to get a referral for a job or learn about jobs that aren’t even posted. Even before that, it will help you get a better idea of the opportunities out there and what might be a good fit. While LinkedIn is a powerful networking tool, there are a bunch of AI- and ML-specific online communities — Discords, Slack channels, and message boards — that are also great to get involved in. You can also keep up with the latest developments in artificial intelligence and machine learning by following industry-leading blogs.

Forward!

If you’ve gotten this far, you’ve already progressed in your understanding of artificial intelligence, how it’s impacting important industries, and how you can get involved. We hope we’ve also provided you with some direction for your further research, both around AI Forward and elsewhere on the web.

If you’d like an easy way to stay up to date with developments in AI and AI education, consider signing up for our newsletter. We’ll never sell your information, and once weekly we’ll deposit a curated selection of the latest news, application tips, and exciting educational programs  — not to mention guest essays written by brilliant students and AI leaders — right into your inbox.

If you’re looking for more structure on your educational journey, consider also signing up for our six-week email course, Fast-Track your AI Future, in which we’ll lead you through the important decisions you need to make to find a degree program or short course that’s right for you.