Chrome Extension
I spent days verifying this list one course at a time, because almost every "free AI courses" article I found online was lying to me in some small way. A course tagged free that actually needs a credit card for a 7-day trial.
A "free" certificate that costs 50 euros. A flagship course that got quietly moved behind a paywall.
I checked every single entry below against its own official page, and I am telling you exactly which ones are 100% free and which ones have a paid catch, so you never have to find out the hard way.
Key Takeaways
What "Free" Actually Means On This List
Before the list, I want to be precise about my own rules, because this is where most round-up articles fall apart.
I only counted a course as free if the actual learning content, videos, readings, notebooks, and assignments, is accessible at zero cost with no credit card required. A course stays on this list even when an optional certificate costs money, but I flag that cost every single time so you know exactly what you are looking at.
I threw out anything that is a free trial of a normally-paid subscription. That is the single most common trick in this space. Coursera's "Grow with Google" AI Essentials course, for example, states plainly that after a trial period the subscription costs $49 USD per month. I am not listing that as free, no matter how many other articles do.
I also want to explain the "audit" pattern you will see repeated across Harvard, MIT-adjacent, and edX-hosted courses. edX's own help center defines it this way:
"As a free audit learner, you will have temporary access to course materials except graded assignments, and you will not earn a certificate at the end of the course."
That means the videos, the readings, and the practice work are yours for free. Only the graded exam and the paper certificate sit behind a paywall. I count that as a real, usable free course, and I tell you the certificate price every time it applies so you can decide for yourself.
Free AI Courses From Big Tech
Google splits its AI education across a few different platforms, and I want to warn you up front: two of its most commonly recommended courses are not actually free. I cover those in the excluded section below. Here is what genuinely is free.
Microsoft
Microsoft's best free AI education actually lives on GitHub, not on its marketing pages, and every one of these is a full open-source curriculum.
On Microsoft Learn itself, all of the following are free, since Microsoft states plainly that "Microsoft Learn training is free and available to anyone":
Amazon And AWS
AWS Skill Builder states outright that "self-paced digital training on AWS Skill Builder is free," and I found a genuinely large free catalog to back that up.
I want to flag one specific trap here. The AWS Certified AI Practitioner exam prep plan mixes free videos with subscription-only practice exams, and AWS's own page says plainly that "some trainings in this learning plan require a subscription." That plan is freemium, not free, and I am leaving it off this list.
IBM
I could not verify that IBM's Coursera-hosted offerings, including the AI Engineering Professional Certificate, are free. Third-party pricing points to roughly $49/month via Coursera Plus, so I am treating those as paid rather than guessing.
NVIDIA
NVIDIA's Deep Learning Institute catalog is mostly paid, running $30 to $90 per course plus optional certificate fees. But a real free tier does exist inside it, mostly shorter, no-code overview courses:
I noticed several of these free NVIDIA courses carry a retirement warning banner on their own pages, so if you click through and see an enrollment cutoff date, that is real and not a display bug.
Meta
Meta's genuinely free AI education is small, and I want to be honest about that rather than padding the list. As far as I can verify, there are exactly three:
All three require only a free Facebook or Meta login and give a free completion badge rather than a paid certificate.
Free AI Courses From Universities
MIT
MIT OpenCourseWare is, in my own testing, the cleanest "free" claim on this entire list. Its own footer states it is "freely sharing knowledge with learners and educators around the world," and OCW has never once offered a certificate of any kind for any course, so there is no paywall to hide behind.
Harvard
Stanford
Stanford's currently-running CS229 and CS231n sections are gated behind a Stanford login, and I want to be upfront about that rather than pretend otherwise. But each course's own page points to genuinely free public archives:
UC Berkeley
Carnegie Mellon
Free AI Courses From AI-Native Education Platforms
DeepLearning.AI
This is where I found the single largest genuinely free catalog on this entire list. DeepLearning.AI splits cleanly into two product lines: short courses at deeplearning.ai/courses/, which are free, and Specializations or Professional Certificates on Coursera, which are paid. Their own course page states plainly, "Learn for Free."
I want to flag one honest caveat: the lesson videos and code notebooks are free, but a paid DeepLearning.AI Pro membership gates some graded assignments and the completion "accomplishment." With that said, here is the free catalog:
That is 41 confirmed free short courses at deeplearning.ai/courses, and their catalog page suggests even more exist beyond what I could fully paginate. For contrast, their Machine Learning Specialization and Deep Learning Specialization live on Coursera and require payment, so do not confuse the two product lines.
Fast.ai
Fast.ai remains, alongside MIT OCW, one of the most honestly free platforms I checked.
I want to flag one important distinction. Fast.ai's homepage now also promotes a newer, separate product from Jeremy Howard called "How to Solve It With Code," and I confirmed directly that this one costs $500 plus an optional $10 per month afterward. That is a completely different product from the classic free fast.ai courses above, so do not let the shared name confuse you.
Hugging Face
Hugging Face's learning hub at huggingface.co/learn is free sitewide, with the LLM Course stating plainly it is "completely free and without ads," and the Agents Course confirming "the certification process is completely free."
Kaggle Learn
Kaggle, owned by Google, states directly that "the courses are provided at no cost to you, and you can now earn certificates." I found no paywall anywhere on the platform.
All 17 courses live at kaggle.com/learn, and unusually for this list, even the completion certificate costs nothing.
Elements of AI
Free AI Courses From Nonprofits, Governments, And AI Labs
freeCodeCamp
OpenAI Academy
I do not think enough people know this exists. OpenAI Academy runs 14 courses across four pathways: Apply AI at Work, Build with AI, Lead AI Adoption, and Teach and Learn with AI. OpenAI's own help center confirms it directly:
"Are Academy courses free? Yes. Academy courses are free. You must have a ChatGPT account to start a course and save your progress."
Courses include AI Foundations, Get Started with Codex, AI for Educators, and AI for College Students. Completion badges are free, though OpenAI is upfront that they "are not certifications and do not guarantee eligibility for a future certification."
Anthropic's Claude Academy
Anthropic runs its own free education platform too, Claude Academy, covering more than 20 structured courses. Their own homepage states it directly:
"Free AI education from Anthropic. Learn how AI works, how to use it with intention, and how to get more from Claude."
Topics span the AI Fluency framework, Claude Code, Claude Cowork, MCP, and sector-specific tracks for educators, nonprofits, and small businesses. Most courses include a free completion badge with no separate paywall.
Anthropic also publishes five free, open-source courses directly on GitHub, no login needed at all: Anthropic API fundamentals, a Prompt Engineering Interactive Tutorial, Real World Prompting, Prompt Evaluations, and Tool Use, all at github.com/anthropics/courses.
Government And NGO Programs
Free Courses On Lead-Generation Platforms
I want to include these because they are genuinely free, while being honest that both platforms will follow up with sales contact after you sign up.
My Honest Review
I went into this expecting to find maybe thirty legitimately free courses. I found well over 150, and that genuinely surprised me.
What struck me most is which organizations are the most honest about pricing. MIT OpenCourseWare, fast.ai, Kaggle Learn, and Hugging Face never once tried to upsell me during this research. Claude Academy and OpenAI Academy were the two biggest surprises, both are real, both are current, and I almost never see either one mentioned in the usual "best free AI courses" articles that keep recycling the same five names.
What frustrated me is how normalized the freemium trap has become. A 7-day trial before a $49 monthly charge gets called "free" constantly, and I had to manually verify pricing pages one at a time to catch it. If you take one thing from my research, take this: always scroll to the actual pricing or FAQ section of a course page before you trust the word "free" in the headline.
If I had to recommend a starting path from everything above, I would send a total beginner to Elements of AI first for the concepts, then Google's Machine Learning Crash Course or Kaggle Learn for hands-on practice, then Anthropic's Claude Academy or DeepLearning.AI's short courses once you are ready to actually build something.
Prompts To Build Your Free AI Learning Path
I use prompts like these to turn a long list like this one into an actual study plan instead of forty open tabs I never return to.
Build a personalized curriculum from this list
I want you to build me a personalized AI learning path using only free courses. My current skill level: [beginner / some coding experience / comfortable with Python / already know ML basics] My goal: [general AI literacy / become an ML engineer / build AI products / use AI tools better at work] Time I can commit: [X hours per week] Timeline: [X weeks] Choose from these providers and pick a sensible order: MIT OpenCourseWare, Harvard CS50 AI, Stanford CS229/CS231n/CS224N, Google Machine Learning Crash Course, Kaggle Learn, Hugging Face courses, fast.ai, DeepLearning.AI short courses, Claude Academy, OpenAI Academy, Elements of AI, Microsoft's GitHub curricula (AI for Beginners, ML for Beginners, Generative AI for Beginners). Give me a week-by-week schedule with specific course names, the order to take them in, and why each one comes before the next. Flag any prerequisites I would be missing.
Turn a lecture into a study guide
I just watched a lecture on [topic, e.g. "backpropagation in neural networks"] from [course name]. Here are my raw notes or a transcript excerpt: [paste your notes or transcript here] Turn this into a structured study guide with: 1. A plain-English summary of the core idea in 3 sentences 2. The 5 most important terms, each defined simply 3. One worked example different from the one in the lecture 4. Three questions I should be able to answer if I actually understood this 5. One real-world application of this concept
Generate a practice quiz after finishing a module
I just completed a module on [topic] from [course name, e.g. "Kaggle Learn's Intro to Deep Learning"]. Write me a 10-question practice quiz mixing multiple choice and short answer, ordered from easiest to hardest. After I answer, grade my responses, explain what I got wrong in plain language, and point me to the specific concept I should review if I got it wrong.
Explain a confusing concept with a different analogy
I'm stuck on this concept from a free AI course and the explanation isn't clicking: [concept]. Here's what I already understand: [what you know] Here's specifically what's confusing me: [the sticking point] Explain it to me using a completely different analogy than whatever is standard for this topic. Then walk through one small, concrete numerical example step by step.
Frequently Asked Questions (FAQs)
Are these AI courses actually completely free, forever?
Every course on this list is free to access right now, verified directly against its own official page. Some include a paid optional certificate, which I flag every time it applies. A small number, mostly at NVIDIA, carry retirement warnings on their own pages, so check the enrollment date before you start.
What does "freemium" mean in the context of AI courses?
I use it to mean a course or platform where the free version is either a time-limited trial of a paid subscription, or a crippled version missing the content you actually need. Google's AI Essentials, at $49 per month after a 7-day trial, is the clearest example I found.
Do I need to pay for a certificate to prove I learned something?
Not necessarily. MIT OpenCourseWare, Kaggle Learn, Hugging Face, Elements of AI's Introduction to AI course, freeCodeCamp, OpenAI Academy, and Claude Academy all offer free certificates or completion credentials. Others, like Harvard's CS50 AI, only charge for the verified version while the course itself stays free.
Which free AI course should I start with as a complete beginner?
I would start with Elements of AI for the concepts with no math or code required, then move to Google's Machine Learning Crash Course or Kaggle Learn's Intro to Machine Learning once you want hands-on practice.
Are Anthropic and OpenAI's free courses worth taking?
Yes, and I think they are underused. Claude Academy covers more than 20 courses on AI fluency and building with Claude, and OpenAI Academy runs 14 courses across four pathways. Both are free and current as of this research.
Final Thoughts
I did not expect a "free courses" round-up to turn into a fact-checking exercise, but that is what this became. The gap between what gets called free online and what is actually free is bigger than I assumed going in, and I would rather hand you a shorter, verified list than a longer one padded with trials and paywalls disguised as courses.
Start with one provider, finish one course, and only then move to the next. A study plan across 150 free courses is not a plan, it is a browser full of guilt. Use the prompts above to cut this list down to the five or six that actually match what you are trying to do.
I am tracking updates to this list as courses change or get retired, along with everything else I use, on promptslove.com.
