Two Must-Have ISTQB AI Certifications
Traditional testing was not built for AI, and the certification world has finally caught up. If you are a tester trying to prove you can test AI without years of hands-on experience, this is for you. There are two new ISTQB AI certifications, and getting in on the ground floor now is the opportunity. The video above breaks down both, and below is the written version covering what each one is, how they differ, and which to pick for your role. These certs did not exist a few months ago, so almost no one on the market has them yet.
Why AI needs its own certification
AI is non-deterministic and data-driven, so it demands a testing approach traditional methods do not cover.
The same input can produce a different output, and the correct response varies with the model and its data. Quality lives inside the training data, and as that data improves, the outputs shift over time. That is the whole point of AI, and it is also why old testing habits break.
On top of that, you deal with bias, hallucinations, and toxicity in the outputs. I explain this in the video. These certifications exist because testing AI is genuinely different, and they give you a vocabulary and method instead of guesswork.
The CT-AI certification: the broad landscape
CT-AI is the broad certification covering how to test AI and machine learning across the full lifecycle.
The first certification is the CT-AI. This one is broader, teaching you how AI works, the impacts of data and models, and the whole AI testing lifecycle. It covers training data quality, bias detection, model performance metrics like precision and recall, drift management in production, pre-trained models, fine-tuning, RAG, and responsible AI ethics.
The exam is 40 multiple-choice questions in 60 minutes, and you need 65 percent to pass. It requires the CTFL foundation plus about six months of experience, and it was released in April 2026. What I learned reviewing the syllabus is that CT-AI is the map of the whole territory, which makes it the right starting point.
The CT-GenAI certification: the deep slice
CT-GenAI goes deep on generative systems: LLMs, RAG pipelines, and non-deterministic content.
If you work with chatbots, LLM apps, and image generation every day, the CT-GenAI is the better fit. It focuses on testing generative systems: unpredictable outputs, content quality, factuality, bias, style, and toxicity. It also covers hallucination versus harmlessness testing, adversarial and prompt testing, building a golden set of reference answers, and wiring generative AI tests into your CI/CD pipeline.
Like the CT-AI, it is 40 questions with a 65 percent pass mark and requires the CTFL foundation first. I get into the details in the video. Knowledge of APIs, JSON, and CI/CD helps. I found that hands-on exposure to generative tools matters more here than anywhere.
Which one should you get
Take CT-AI for machine learning and predictions, CT-GenAI for chatbots and LLM apps, and both for the strongest resume.
The choice comes down to your work. If you test machine learning models, data pipelines, and predictions, CT-AI is best. If you test chatbots, LLM apps, RAG, and image generation, go CT-GenAI. New to AI testing and want the broad foundation? Take CT-AI first. Living in generative AI daily? CT-GenAI. For the best resume, do the CTFL foundation, then CT-AI, then CT-GenAI.
I share my exam approach in the video. A few tips from taking many of these: study the syllabus and glossary until the terms are second nature, read each question twice, and watch the verbs like “best,” “except,” and “never.” I recommend five or six timed practice exams so your pacing is solid, and pair the certification with real hands-on work so the knowledge sticks.
The takeaway
The two new ISTQB AI certifications give you a credible way to prove you can test AI. CT-AI is the broad lifecycle certification for machine learning, data, and predictions. CT-GenAI is the deep slice for LLMs, RAG, and generative content. Both require the CTFL foundation, both are 40 questions at 65 percent to pass, and both are rare enough right now to put you ahead of the pack. Get the foundation, add CT-AI, then specialize with CT-GenAI, and back it all with real practice.
Watch the full comparison in my video on the two ISTQB AI certifications, and I cover the pick-the-right-one guidance further in the video. Here is my question for the comments: are you going CT-AI, CT-GenAI, or both? Subscribe if you want more on QA certifications and careers.
