The Best AI Courses to Take in 2026: A Comprehensive Guide

I rebuilt this list from scratch in September 2026, and the first thing I noticed is that "AI course" now describes three different products.

There are short courses that teach you to use ChatGPT, Claude, or Copilot at work. There are engineering tracks that teach you to build on top of models someone else trained. And there is graduate material for people who want to understand the models themselves.

Most rankings put all three in a single list, which is how a 1-hour ChatGPT course ends up ranked against a Berkeley reinforcement learning class. I've split them into 8 categories instead, starting with the ones most people actually need.

If you want to build AI applications, our guide to the best AI engineering courses goes much deeper than the section here. You can also see our lists of the best free AI courses and the best machine learning courses.

Course ratings, review counts and learner numbers last verified 24 September 2026.

Best AI Courses Comparison Table

Every DataCamp course and track in this article is listed below with its live rating and learner count. Courses from other providers are covered in the sections themselves.

Best AI Courses For Beginners

The best AI course for a complete beginner explains what AI is before it asks you to use it, and asks nothing technical of you along the way.

Beginners also get sold coding courses they don't need. Understanding what a large language model (LLM) does, and where it fails, takes an afternoon, not a Python course.

1. Understanding Artificial Intelligence - DataCamp

Our conceptual starting point, and the one I recommend to anyone who wants the vocabulary before the tools.

• Level: Beginner, no coding
• Time: 2 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Anyone who needs to follow AI conversations at work before using AI themselves

Four chapters cover what AI is, which tasks it solves, how organizations put it to work, and its human side, across 16 videos and 49 exercises. It's taught by Iván Palomares Carrascosa.

At 423K learners, it's one of the most-taken courses in our AI catalog. That volume matters for a beginner course: every confusing exercise has been reported and fixed many times over.

2. AI Fundamentals - DataCamp

The structured version of this section: one track, a set order, and a certification at the end.

• Level: Beginner, no coding
• Time: 9 hours
• Cost: Included with a subscription, from $14/month billed annually
• Best for: People who want one path to follow and a credential to show for it

Across 5 courses, it covers AI for work, machine learning, LLMs, generative AI, and AI ethics, all without code.

It's also the preparation track for our AI Fundamentals Certification, which I cover in the certification section below. If you take the track, you can skip the standalone courses it contains.

3. Understanding Machine Learning - DataCamp

The two-hour answer to what your data team means when they say "we trained a model."

• Level: Beginner, no coding
• Time: 2 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Managers and analysts who work next to machine learning without building it

Three chapters cover what machine learning is, the main model types, and deep learning, across 12 videos and 36 exercises. Hadrien Lacroix, Sara Billen, and Lis Sulmont teach it.

It's narrower than the course above. Take it second, or take it first if machine learning, not chatbots, is what your job keeps running into.

4. AI Capabilities and Limitations - Anthropic

A free course on why AI tools behave the way they do, and the best free beginner course I found.

• Level: Beginner, no technical background needed
• Time: 13 lessons, about 3.5 hours
• Cost: Free, with a certificate
• Best for: Anyone who has been surprised by a confident wrong answer from a chatbot

It's built around four properties of AI systems: next-token prediction, knowledge, working memory, and steerability. Each section pairs a short explanation with a hands-on exercise, so you see where a model's limits are rather than read about them.

Anthropic uses the same material to onboard its own staff. The examples use Claude, but the four properties apply to every chatbot you'll use.

Our pick: Start with our Understanding Artificial Intelligence for concepts in 2 hours, or the AI Fundamentals track if you want a full path and a certification. If you want it free, take Anthropic's AI Capabilities and Limitations.

Best AI Courses For Using AI At Work

The best course for using AI at work teaches habits that survive the next model release: how to split a task between you and the tool, how to give a prompt enough context, and how to check what comes back.

Tool-specific courses come second, and only for the tool your employer actually pays for. A Copilot course is wasted on a company that runs on Claude.

1. Introduction to AI for Work - DataCamp

Our best overall pick in this article, and the most recently updated course in it.

• Level: Beginner, no experience needed
• Time: 2 hours, about 2.5 hours in the AI Tutor version
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Anyone who wants to be productive with AI this week

The AI Tutor version adapts each lesson to your role and level as you go. It's taught by James Chapman, who leads our AI curriculum, and Yusuf Saber, our Chief AI Officer, who founded the AI-native learning platform Optima before we acquired it in 2025.

The exercises run against a real chatbot inside the course: you'll draft social media posts, run a business analysis, and practice a four-part prompting framework built on the ask, requirements, context, and examples.

The last chapter covers hallucinations, bias, and data privacy. I'd argue it's the most useful 30 minutes in the course, because it's what separates people who use AI from people who trust it too much.

2. Introduction to ChatGPT - DataCamp

A 1-hour playbook for the chatbot most people already have open.

• Level: Beginner
• Time: 1 hour
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Regular ChatGPT users who want better, more reliable answers

Two chapters, Getting to Know ChatGPT and The ChatGPT Playbook, cover better prompts, checking responses, and safe use, across 4 videos and 18 exercises. Joe Franklin teaches it.

One editorial note: our older Understanding ChatGPT course has around 10x the learners, but its content was last updated in September 2025. For a product that changes this often, I picked the course updated in July 2026.

3. Understanding Prompt Engineering - DataCamp

The course that makes every other tool on this list work better.

• Level: Beginner
• Time: 1 hour
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Anyone whose AI answers keep coming back generic

Three chapters move from prompting basics through strategies and techniques to advanced prompt engineering, across 11 videos and 29 exercises. Alex Banks teaches it.

The exercises use ChatGPT, but the techniques carry over to Claude, Gemini, and Copilot. It was last updated in August 2026.

4. Introduction to Microsoft Copilot - DataCamp

The course for anyone whose company runs on Microsoft 365.

• Level: Beginner
• Time: 1 hour
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Microsoft 365 users with a Copilot license

Three chapters cover Copilot and Microsoft 365 Copilot, working with it across the Microsoft apps, and best practices for responsible use, across 11 videos and 34 exercises. Shreya Vashist teaches it.

It's the oldest update in this section, from February 2026. Which Copilot features you see also depends on your company's license, so some exercises may not match your screen.

5. AI Fluency: Framework & Foundations - Anthropic

A free, tool-agnostic course from Anthropic, and the best framework here for thinking about AI collaboration.

• Level: Beginner
• Time: Self-paced
• Cost: Free, with a certificate after a final assessment
• Best for: People who want a mental model that works with any AI tool

It's built around a 4D framework: Delegation, Description, Discernment, and Diligence. Anthropic developed it with professors Rick Dakan of Ringling College of Art and Design and Joseph Feller of University College Cork, and it runs on Anthropic Academy and Coursera.

It's more abstract than the courses above, with fewer hands-on exercises. It works best paired with a tool course rather than instead of one.

Our pick: Take our Introduction to AI for Work first, then the course for your company's tool. If you want a framework rather than a tool, add Anthropic's AI Fluency, which is free.

Best AI Courses For Finance, Marketing, HR, And Sales

Role-specific AI courses are worth taking once the basics are in place, because the prompts that save a finance analyst an hour look nothing like the ones that help a recruiter.

Ours are new, and all four were updated between May and June 2026. Their learner counts are small next to the courses above, so judge them on the syllabus rather than the numbers.

1. AI for Finance - DataCamp

The role course with the widest pull so far, and the one I'd hand a finance team first.

• Level: Beginner
• Time: 3 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Finance and FP&A professionals

Three chapters move from prompting fundamentals for finance to deeper analysis and then to designing finance workflows with AI, across 10 videos and 30 exercises. Glenn Hopper teaches it, and it's built around Microsoft Copilot.

2. AI for Marketing - DataCamp

Campaign work, from planning to execution.

• Level: Beginner
• Time: 3 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Marketers who plan and run campaigns

The chapters cover AI for campaign planning, for campaign execution, and for building custom assistants that handle recurring marketing work, across 10 videos and 29 exercises. Adel Nehme teaches it.

3. AI for Human Resources - DataCamp

Recruiting and people operations, with fairness built into the first chapter.

• Level: Beginner
• Time: 3 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: HR generalists, recruiters, and people ops teams

The three chapters cover faster and fairer recruitment, people operations, and scaling HR support with AI agents, across 11 videos and 32 exercises. Christopher Klaus teaches it.

Hiring is where AI bias has real consequences for real people, so I'm glad the course opens there rather than treating it as an afterthought.

4. AI for Sales - DataCamp

Prospecting, emails, and CRM work, done AI-first.

• Level: Beginner
• Time: 2 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Account executives and sales development reps

Chapters cover prospecting, conversions, and sales coaching, across 9 videos and 26 exercises. Jody Geiger teaches it.

The final chapter uses Copilot Studio, so it's most useful to teams already on Microsoft tools.

5. AI Business Fundamentals - DataCamp

The track for the people who decide where AI goes in a company, rather than the people using it daily.

• Level: Beginner
• Time: 10 hours
• Cost: Included with a subscription
• Best for: Managers and leaders deciding on AI projects

It covers how to scope AI opportunities, build a proof of concept, implement solutions, and put an AI strategy together.

Our pick: Take the course for your role. If yours isn't listed, analysts can take AI for Data Analysts, and managers should take the AI Business Fundamentals track.

Best Free AI Courses

The best free AI course depends on what you want from it: Anthropic's AI Capabilities and Limitations for understanding, the Hugging Face LLM Course or Kaggle's AI Agents Intensive for building, and MIT's 6.S191 for deep learning.

This is also the category where free material most often beats paid material, and I'd rather say that plainly. Our full list is in our guide to the best free AI courses.

1. MIT 6.S191 Introduction to Deep Learning - MIT

The most current free deep learning course, rebuilt every year.

• Level: Intermediate
• Time: A lecture series plus labs, self-paced
• Cost: Free
• Best for: Anyone who wants a fast, current grounding in deep learning

Alexander and Ava Amini released the 2026 lectures on YouTube between March and May 2026, and the labs run in Google Colab, so there's nothing to install. MIT open-sources all materials after each edition.

The course site asks for elementary linear algebra and calculus. That's a light bar, but it's a bar, so it isn't the place to start if Understanding Artificial Intelligence was new to you.

2. LLM Course - Hugging Face

The reference course for the open-source model stack, written by the people who build it.

• Level: Intermediate, strong Python needed
• Time: 12 chapters, self-paced
• Cost: Free
• Best for: Developers working with open-weight models

Chapters 1 to 4 cover the Transformers library, 5 to 8 cover the Datasets and Tokenizers libraries and classic language tasks, chapter 9 covers demos, and chapters 10 to 12 cover fine-tuning, curating datasets, and building reasoning models.

Hugging Face recommends an introductory deep learning course first, such as MIT's above. The course introduction also says there's no certification for the course as a whole.

3. 5-Day AI Agents Intensive - Kaggle and Google

A free, self-paced agents course, and by attendance one of the largest technical courses ever run.

• Level: Intermediate, Python needed
• Time: 5 days of material, self-paced
• Cost: Free
• Best for: Developers who want to build their first AI agent

Google's machine learning researchers and engineers built it, and it ran live from November 10 to 14, 2025, before moving to a self-paced Kaggle Learn guide. It covers the core parts of an agent: models, tools, orchestration, memory, and evaluation.

According to KDnuggets, more than 1.5 million people signed up for the live edition and filed more than 11,000 capstone submissions. Google ran a follow-up edition focused on vibe coding in June 2026.

4. AI Capabilities and Limitations - Anthropic

Our free pick for complete beginners, for the reasons in the beginner section above.

• Level: Beginner
• Time: About 3.5 hours
• Cost: Free
• Best for: Non-technical learners

It's the only course in this section that needs no code and no math at all.

5. Claude 101 - DataCamp and Anthropic Academy

Built by Anthropic Academy and hosted on our platform, and free in full.

• Level: Beginner
• Time: 2 hours
• Cost: Free
• Best for: Teams whose company has rolled out Claude

Four chapters cover meeting Claude, organizing your work and knowledge in it, extending what it can reach, and a final chapter that puts it together, across 20 exercises.

It's the only course on our platform in this article that's free in full. Every other course has its first chapter free, not the whole course. Claude's feature set also changes quickly, so expect some screens to look different from the course within months.

Our pick: Take Anthropic's AI Capabilities and Limitations if you have never studied AI, MIT 6.S191 if you want deep learning, and the Hugging Face LLM Course or Kaggle's AI Agents Intensive if you want to build.

Best AI Courses With Certification

An AI certification is worth paying for when you need to show an employer something before you have a portfolio to show them.

Check what kind it is before you book, because a completion certificate, a proctored exam, and a credential tied to one vendor's cloud send three different signals.

While verifying this section, I also found that one of the most-recommended entry certifications, Microsoft's AI-900 exam, retired on June 30, 2026. Several roundups still recommend it.

1. AI Fundamentals Certification - DataCamp

Our no-code certification, and the one that pairs with the AI Fundamentals track.

• Level: Beginner, no coding
• Time: One 30-question timed exam, up to 1 hour
• Cost: Included with a Premium subscription
• Best for: Non-technical professionals who want a first AI credential

You can retake the exam every 30 days, and our Fundamentals Certifications don't expire. The AI Fundamentals track is the preparation route.

It's a DataCamp credential, so it carries less name recognition with hiring managers than an AWS or Google certification.

2. AI Engineer for Developers Associate - DataCamp

A two-stage certification that tests whether you can build, not just recall.

• Level: Associate
• Time: A timed exam, then a 4-hour practical exam with 3 tasks
• Cost: Included with a Premium subscription
• Best for: Developers moving into AI engineering

It covers AI models, prompt engineering, implementing AI solutions, and software engineering principles, including building chatbots. The practical exam is graded automatically, you get two attempts, and you have 30 days from registration to finish both parts.

Our career certifications are valid for 2 years. The Associate AI Engineer for Developers track is the preparation route.

3. AWS Certified AI Practitioner (AIF-C01) - AWS

The strongest vendor credential for people who use AI on AWS rather than build it.

• Level: Foundational
• Time: 90 minutes, 65 questions
• Cost: $100 per attempt
• Best for: Business analysts, product managers, and IT staff in AWS organizations

Of the 65 questions, 50 are scored, and you need 700 out of 1,000 to pass. It's valid for 3 years and has no prerequisites.

It tests AWS services by name, including Amazon Bedrock and SageMaker, so much of what you study only applies inside AWS.

4. Generative AI Leader - Google Cloud

Google's least technical certification, aimed at the people sponsoring AI projects.

• Level: Beginner, no coding
• Time: 90 minutes, 50 to 60 multiple-choice questions
• Cost: $99 per attempt
• Best for: Managers and business-side professionals

Google Cloud says the certification is for anyone in any job role, with or without hands-on technical experience. It's valid for 3 years.

Expect a lot of Google product names on the exam. It teaches generative AI through Google Cloud's portfolio, so it's a weaker fit if your company doesn't use it.

5. Azure AI Fundamentals (AI-901) - Microsoft

The replacement for AI-900, and no longer a no-code exam.

• Level: Beginner
• Time: Exam only
• Cost: Paid exam, pricing varies by region
• Best for: People working in Microsoft Azure environments

The AI-900 exam retired on June 30, 2026. The certification, Microsoft Certified: Azure AI Fundamentals, continues through AI-901, which focuses on building AI apps and agents in Microsoft Foundry.

Microsoft now recommends some familiarity with Python syntax, which AI-900 never asked for. If you already hold the certification through AI-900, you keep it.

Our pick: Take our AI Fundamentals Certification if you're non-technical, AWS Certified AI Practitioner if your employer runs on AWS, and our AI Engineer for Developers Associate if you build software.

Best AI Courses For AI Engineering

AI engineering courses teach you to build products on top of existing models: calling APIs, retrieval-augmented generation (RAG), agents, and evaluation. None of it requires training a model yourself.

This section is deliberately short. Our guide to the best AI engineering courses covers agent frameworks, fine-tuning, and LLMOps in much more depth.

1. Associate AI Engineer for Developers - DataCamp

Our track for developers adding AI to real software, and the best starting point if you know Python but have never touched machine learning.

• Level: Beginner in AI, Python required
• Time: 29 hours
• Cost: Included with a subscription, from $14/month billed annually
• Best for: Developers who want to build AI features

It covers the OpenAI API, Hugging Face's models and datasets, LangChain, and the Pinecone vector database, and you'll build chatbots, recommendation engines, and semantic search along the way.

It also covers the unglamorous parts: rate limits, error handling, and structuring model outputs so they don't break the code around them. It prepares you for the AI Engineer for Developers Associate certification.

2. Associate AI Engineer for Data Scientists - DataCamp

The same destination, starting from the data science side.

• Level: Intermediate
• Time: 40 hours
• Cost: Included with a subscription
• Best for: Data scientists moving into AI engineering

It focuses on training and fine-tuning models, including LLMs, for production. The matching certification has two timed exams and a practical exam that must be completed in Python.

3. Working with the OpenAI API - DataCamp

The shortest route to your first working API call.

• Level: Beginner in AI, Python required
• Time: 3 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Developers who want to try an LLM API before committing to a track

Three chapters cover the API itself, prompting OpenAI models, and building multi-turn conversations, across 9 videos and 29 exercises. James Chapman and Eduardo Oliveira teach it.

It's OpenAI-specific. The concepts carry over to Anthropic's and Google's APIs, but the code won't.

4. LLM Course - Hugging Face

Covered in the free section above, and the natural next step once the OpenAI API feels familiar.

• Level: Intermediate
• Time: 12 chapters
• Cost: Free
• Best for: Developers moving to open-weight models

Take it when you need to run or fine-tune models yourself rather than call them through an API.

Our pick: Take our Associate AI Engineer for Developers track if you come from software, or the Associate AI Engineer for Data Scientists track if you come from data science.

Best Machine Learning And Deep Learning Courses

Machine learning courses teach you to train models on your own data, which is a different job from calling a model someone else trained.

You'll need Python and some statistics first. The payoff is being able to build the models that the AI engineering section above treats as a black box.

1. Machine Learning Scientist in Python - DataCamp

The longest and most demanding path on this list.

• Level: Advanced
• Time: 85 hours
• Cost: Included with a subscription, from $14/month billed annually
• Best for: People moving from analysis into machine learning

It covers supervised learning, unsupervised learning, and deep learning in Python. At 85 hours, it's a months-long commitment for most people, not a weekend.

2. Supervised Learning with scikit-learn - DataCamp

Where most people's machine learning actually starts.

• Level: Intermediate
• Time: 4 hours
• Cost: Chapter 1 free, full access from $14/month billed annually
• Best for: Python users who want to build their first predictive model

Four chapters cover classification, regression, tuning your model, and preprocessing with pipelines, across 15 videos and 54 exercises. George Boorman teaches it.

If you only take one course in this section, take this one. Most real-world machine learning is still classification and regression on tables.

3. Deep Learning in Python - DataCamp

Our PyTorch track, and the step after scikit-learn.

• Level: Advanced
• Time: 18 hours
• Cost: Included with a subscription
• Best for: People heading toward research or machine learning engineering

It teaches you to build neural networks in PyTorch for different data types, starting with our Introduction to Deep Learning with PyTorch course.

You'll want comfort with NumPy and some linear algebra before you start.

4. PyTorch for Deep Learning Professional Certificate - DeepLearning.AI, Coursera

The newest structured PyTorch program from outside our catalog, launched in October 2025.

• Level: Intermediate
• Time: 3 courses, about 2 months at 10 hours a week, according to Coursera
• Cost: Coursera subscription
• Best for: Learners who want PyTorch taught through to deployment

Laurence Moroney teaches it across three courses: PyTorch fundamentals, techniques and ecosystem tools, and advanced architectures and deployment. It covers ResNets, transformers, and diffusion models, then ONNX, MLflow, pruning, and quantization.

It overlaps heavily with our Deep Learning in Python track, so pick one. Choose this one if you want the DeepLearning.AI name on the certificate.

5. MIT 6.S191 Introduction to Deep Learning - MIT

Covered in the free section above, and the fastest current overview before a longer program.

• Level: Intermediate
• Time: A lecture series plus labs
• Cost: Free
• Best for: Anyone who wants the big picture of deep learning first

Watch the 2026 lectures before starting a PyTorch track, and the track will make more sense.

Our pick: Start with Supervised Learning with scikit-learn, then move to the Machine Learning Scientist in Python track for depth or Deep Learning in Python for neural networks.

Best Advanced AI Courses For Research

Advanced AI courses are mostly graduate courses published for free, and they assume you can read a paper and implement it.

I've kept our own material to one entry here, on purpose. For research-level AI, the best courses come from the labs and universities doing the research.

1. Machine Learning Engineer - DataCamp

Our track for getting models into production and keeping them there.

• Level: Advanced, assumes Python and model training
• Time: 44 hours
• Cost: Included with a subscription
• Best for: Data scientists whose models need to survive production

Its 12 courses cover MLOps concepts, Docker, MLflow, data pipelines, CI/CD for machine learning, and monitoring models in production.

It's the engineering side of advanced work, not the research side, which is why the rest of this section is external.

2. CS25 Transformers United - Stanford

A public seminar where researchers present their own work on transformers.

• Level: Advanced
• Time: A weekly seminar over one quarter
• Cost: Free to audit
• Best for: People tracking where transformer research is heading

Past speakers include Geoffrey Hinton, Ashish Vaswani, and Andrej Karpathy, and the sixth edition ran in spring 2026. Anyone can join the Zoom livestreams without signing up, and recordings go to YouTube.

The only homework for Stanford students is attending, and there's no certificate for public viewers.

3. 11-711 Advanced NLP - Carnegie Mellon

The most current graduate NLP course you can follow for free.

• Level: Advanced
• Time: A full semester of lectures
• Cost: Free lectures and readings
• Best for: Engineers and researchers who want LLM internals

The spring 2026 lectures are on YouTube, including sessions by Sean Welleck on reinforcement learning for LLMs and on parallel and distributed training. The course site lists the readings for each class.

The assignments and grading are for CMU students, so outside learners get the lectures and papers, not feedback.

4. CS285 Deep Reinforcement Learning - UC Berkeley

The standard graduate introduction to deep reinforcement learning.

• Level: Advanced
• Time: A full semester of recorded lectures
• Cost: Free
• Best for: Researchers and engineers working on reinforcement learning

Sergey Levine has taught it for years, and the spring 2026 lectures are on YouTube. It assumes a machine learning foundation, including Markov decision processes, and it's math-heavy.

Reinforcement learning also underpins how today's LLMs are post-trained, which makes this course more relevant to language model work than its title suggests.

5. OMSCS, Artificial Intelligence Specialization - Georgia Tech

The best value accredited AI degree available online.

• Level: Graduate
• Time: A 30-credit-hour master's degree, typically 2 to 3 years part-time
• Cost: About $236 per credit hour, roughly $7,000 in tuition plus per-semester fees
• Best for: Career-track learners who want a real graduate credential

Once fees are included, 2026 estimates for the full degree land between about $9,000 and $9,400, depending on how many semesters you take.

Admission is required, and it's a degree, not a course. It sits here because nothing else on this list gives you an accredited master's.

Our pick: Take our Machine Learning Engineer track if your goal is production, CMU's Advanced NLP if it's language models, and Georgia Tech's OMSCS if you need the degree.

How We Chose These Courses

The ratings, review counts, and learner numbers for our own courses are pulled from live course data, on the date shown in the freshness stamp at the top of this page. For tracks, the rating is a review-weighted average of the courses the track contains, since a track page alone collects too few reviews to be a useful signal.

Any course with no content update or new recording in the past 18 months didn't make the cut, whether it was ours or someone else's. A few familiar names that still appear in most roundups fell out as a result.

External courses were checked against the provider's own pages or recent coverage. Where a free or competing course is the better choice, we included it and said so in that section.

Final Thoughts

For most people, our Introduction to AI for Work is where I'd start, because it teaches the habits that make every AI tool more useful, in 2 hours.

From there, the path splits. If you use AI, add the course for your tool or your role. If you want to build with it, move to our Associate AI Engineer for Developers track.

A few caveats about this list. The certification market is moving fast, as Microsoft's AI-900 retirement shows, so check the provider's page before you book an exam. And tool courses date quickly: expect Claude and Copilot screens to change within months of any course being recorded.

If you want to start today and follow one path, I recommend the AI Fundamentals track, which covers the beginner section above in 9 hours and prepares you for a certification at the end.