Free AI Courses From Amazon: Are They Worth Taking?

In January 2024, we first covered “AI Ready,” a new line of online artificial intelligence literacy courses offered by Amazon Web Services for free to two million people around the world. AI Ready was the first AI training program we at OnlineEducation.com had found that was offered for free and at scale to a massive global student audience.

At the time, we also learned from a November 2023 AWS survey of hiring managers released with the AI Ready launch that United States employers were paying as much as 47 percent in salary premiums to employees who demonstrate sufficient AI skills from the research for our article “How Much Money Can AI-Skilled Workers Make?”

Moreover, a February 2024 employee survey by Indeed replicated Amazon’s results precisely. Indeed’s “2024 Workforce Insights Report,” which polled more than 5,000 U.S. workers, found the same 47 percent salary boost for American employees with AI skills. The poll also found an average annual salary for these workers approaching a whopping $175,000.

Furthermore, a new survey by the Adecco Group’s tech training firm, General Assembly, revealed several interesting insights about the market for AI skills in Australia, Canada, France, Singapore and the United States. Fifty-two percent of the poll’s sample of 1,000 human resources managers hiring for software engineering, UX, and data analytics jobs said they’re spending more than $10,000 to fill roles requiring AI skills.

In fact, two-thirds of the time, their companies agree to pay what job candidates ask for during the hiring process. Additionally, to acquire AI talent, more than half (52 percent) are hiring extra HR staff, and 53 percent are reducing their open positions’ traditional education requirements.

As we researched the needs for AI skills within the business community, the extent to which America’s general population also needs AI training became apparent as well. On January 23 EdSurge released a remarkable podcast interview with Susan Gonzales, a former Facebook executive who manages the San Francisco-based advocacy organization AIandYou which promotes AI literacy. Gonzalez also serves as a workgroup co-chair of President Biden’s National AI Advisory Committee (NAIAC).

She told EdSurge’s Editor Jeffrey Young that polling and opinion research conducted across America late in 2023 for NAIAC’s Education and Awareness Workgroup revealed that the levels of awareness about fundamental artificial intelligence concepts are very low, particularly among Americans in underrepresented communities.

“Most people still believe that AI is coming—someday—and that robots are going to take over,” says Gonzalez. “So we have a long way to go in educating.”

In the sort of environment Gonzalez describes, free and easily available courses like Amazon’s could make a tremendous difference. However, some folks, like the Wall Street Journal’s reporter Sebastian Herrera, have raised concerns about Amazon’s new offerings.

Amazon Web Services (AWS) is Amazon’s subsidiary that provides on-demand cloud computing platforms to companies and governments; the subsidiary earned $80 billion in revenue in 2022. In this video report, Herrera argues that AWS currently competes for scarce talent that possesses artificial intelligence skills against rivals that have invested billions of dollars in AI, including Microsoft and Google. Because the AI Ready courses are run on Amazon Web Services AI platforms, one could argue that by enabling up to two million people to become familiar with Amazon’s systems suddenly, these ostensible “courses” could also be used by the company as recruiting tools.

Herrera also argues that the courses could benefit Amazon’s partner companies. Such firms include enterprise customers that purchase the web, cloud and generative AI services provided by AWS—and those customers are also having a lot of trouble finding and hiring scarce talent who already possess AI skills.

Given Gonzalez’s dire assessment of AI literacy across the United States combined with the concerns about the AI Ready program raised by the WSJ’s Herrera, it’s reasonable to wonder about these courses’ educational quality. With not that many high-quality AI training options yet available for free as of July 2024, are Amazon’s courses worth taking? Do they deliver sufficient value that’s worth the opportunity costs of our readers’ investments in time? Or are these courses mainly recruiting, promotional, and public relations tools designed to boost Amazon’s profits?

We decided to answer these questions by taking a few of Amazon’s AI courses. What we learned surprised us. It turns out that Amazon Web Services does in fact offer outstanding training in artificial intelligence fundamentals for free to the global public at scale.

However, AWS doesn’t offer those classes within the AI Ready courseware program. And unless one learns about Amazon’s best-quality AI training alternatives through an article like this one, it’s unclear how one would ever come across those courses.

Available through the AWS Educate platform alongside courses in cloud computing skills, AWS offers this brief, self-paced, foundational-level AI Ready digital training course for people just starting out with artificial intelligence.

The Introduction to Generative Artificial Intelligence course requires no prior understanding of AI concepts. It might also amount to the fastest and most concise educational AI presentation ever developed, given that Amazon tells learners to budget only about 45 minutes for the entire course: AWS roughly allocates 20 minutes for the learning content’s presentation, followed by 20 minutes to take a multiple-choice quiz.

There are only four topics:

1. A definition of generative artificial intelligence and a comparison of how GenAI differs from older, traditional AI
2. A definition of foundation models and a comparison of how foundation models differ from traditional models
3. A brief explanation of prompt engineering: how AI prompts work, followed by an example of a timesaving prompt that automatically writes lines of software code
4. Examples of generative AI use cases for six industry sectors: healthcare, life sciences, financial services, manufacturing, retail, and media/entertainment

Should a potential learner spend 45 minutes with this courseware? That depends. A business user with absolutely no familiarity with very basic AI concepts will probably find this course worthwhile—so long as they can tolerate 20 minutes worth of the courseware’s rudimentary, bare-bones production quality.

For example, Amazon doesn’t use any photos or graphic images as illustrations in this courseware’s slides—only the simplest of drawings appear along with common symbols. Strangely enough, Amazon also uses a grating, poor-quality text-to-speech voice to narrate the slides. That’s a curious choice, given that AWS could have selected one of the high-quality TTS voices used on the Amazon Echo smart speakers and produced by its own text-to-speech software subsidiary, Ivona. Or better yet, AWS could have opted for live recordings by voiceover professionals.

Amazon also relies on learners to silently read 14 tedious college textbook-style passages within the courseware that appear without any narration at all. These passages surface within slides that explain the three foundation models and six use cases. AWS even presents five blatant advertisements this way that pitch the “benefits of using AWS for your generative AI solutions.” And yes, questions about this textbook material do appear on the quiz—and a sufficient score earns a digital badge.

All in all, few people who sit through this course will consider it to have been a worthwhile experience. And that’s especially true of students used to the much higher-quality courseware routinely used by all the top universities we regularly cover here on OnlineEducation.com and our family of college- and university-focused websites.

The risk for Amazon is that people who take this course might react so negatively that they won’t want to take any of the better-quality courses available from AWS, such as the terrific video lecture class we describe coming up in our concluding section at the end of this article.

Intended for developers, data scientists, and prompt engineers without much formal training, this four-hour, self-paced digital course on prompt engineering provides a much higher-quality presentation than the courseware in the previous Introduction to Generative Artificial Intelligence class. However, be forewarned that this course’s lessons take longer to complete and are much more complex and technically involved than Amazon Web Services’ descriptions would lead potential students to believe.

The basic prompt engineering techniques presented by the first segment of this course help students return more effective results from their AI applications. In AI lingo, a “shot” is an example, so the course first talks about zero-shot techniques that don’t show a large language model any examples. The presentation continues with an explanation of how to pre-train and fine-tune the LLM’s results by showing it examples using few-shot and chain-of-thought techniques, which will usually return better quality results than the zero-shot methods.

However, it’s unclear why Amazon decided to offer this class as a single four-hour session instead of dividing the seven lessons between a pair of two-hour sessions. Prompt engineering is an important skill for business users as well as developers, but most business users would rarely—if ever—use some of the cutting-edge techniques presented later in this course.

Those advanced techniques that should have been reserved for a second class include Tree of Thoughts, Retrieval-Augmented Generation (RAG), Automatic Reasoning and Tool-Use (ART), and Reasoning and Acting (ReAct). Scholarly journal articles accompanying the lessons indicate that most of these methods were developed only a few months ago by researchers at firms like OpenAI and Google along with several universities. Presentations about specific models, such as Amazon’s Titan, Anthropic’s Claude, and AI21 Labs’ Jurassic-2, also should have been reserved for a second course as well.

Although the course doesn’t require a quiz, the learning management system won’t award a certificate to a graduate unless they complete every section. We also encountered repeated difficulties logging in to this course on the AWS Builder platform. That’s because the login screen frequently failed to present captcha images, completely blocking access to our previously completed course, including its content and certificate. Unfortunately, a detailed incident report emailed to the AWS help desk returned an email reply two days later indicating that the specialist failed to understand the issues we had reported.

Fortunately, AWS now offers a new AI training course that’s far superior to the digital courses available through AI Ready, such as the two problematic courses we discuss above but cannot recommend.

Generative AI Foundations on AWS instead offers a self-paced YouTube playlist of seven free video lectures about generative AI foundations by Emily Webber, an astute and personable machine learning expert with six years of experience at Amazon and a master’s degree in computational analysis from the University of Chicago. And anyone who spends only five minutes watching excerpts from her courses—in which she presents clear and engaging explanations of complex topics paired with simple and vivid graphics—will probably cross the two AI Ready courses we reviewed above off their to-do list.

Here are her topics for each class:

1. Introduction to Foundation Models
2. Picking the Right Foundation Model
3. Prompt Engineering and Fine-Tuning
4. Pretraining a New Foundation Model
5. Preparing Data and Training at Scale
6. Reinforcement Learning With Human Feedback
7. Deploying a Foundation Model

Webber also hosts a video podcast on Twitch called Build on Generative AI. Her talk show presents AI tips, tricks, and guidance in a much more informal and less structured format than that of the Generative AI Foundations series of classes.

Judging by the surprisingly low view counts so far on YouTube, Webber’s fantastic free lectures are arguably the best-kept secret in AI training right now—and they earned our enthusiastic recommendation.

However, it’s not clear how potential students attracted by all the publicity about the AI Ready program would ever find her courses. We didn’t notice any links from the AI Ready websites to her playlist, and aside from a single report in addition to this article, we couldn’t find much press coverage about Webber’s class.