X-GenAI
← All posts
AI certification6 min read

AI certifications in 2026: what's changed, and what's worth your time

AWS retired its Machine Learning Specialty, Microsoft replaced its Azure AI Engineer exam with one built around agents, and the EU's AI literacy rule is driving demand for basics. A status update.

The AI certification landscape has shifted more in the past year than in the five before it. The big cloud providers are retiring exams built around classic machine learning and replacing them with ones about generative AI and agents. At the same time, regulation has created demand for basic AI literacy training. Here's what changed and how to think about it.

What changed

  • AWS. The AWS Certified Machine Learning, Specialty exam has been retired. The last day to sit it was 31 March 2026. Existing holders keep it for three years from the date they earned it. AWS points people to the Machine Learning Engineer, Associate and the AI Practitioner exams, and in late 2025 opened a beta for a new Generative AI Developer, Professional certification focused on building with foundation models.
  • Microsoft. The Azure AI Engineer Associate (exam AI-102) is due to retire on 30 June 2026. Its replacement, AI-103, leads to a new Azure AI Apps and Agents Developer Associate credential, with much more weight on building agentic applications.
  • Regulation. Article 4 of the EU AI Act has required organisations that use AI to ensure sufficient AI literacy among their staff since 2 February 2025. That has pushed a lot of employers toward short fundamentals courses.
What changed in the major cloud AI certifications.

AWS

Machine Learning, Specialty

Retired March 2026

Machine Learning Engineer, Associate

Generative AI Developer, Professional

Beta from late 2025

Azure

AI Engineer, AI‑102

Retiring June 2026

AI Apps and Agents Developer, AI‑103

Figure 1. What changed in the major cloud AI certifications.

What it means

The direction is clear: certifications are following the work. Exams about training and tuning traditional models are giving way to exams about deploying foundation models, retrieval and agents. If you're studying from guides more than a year old, check that the exam still exists.

Our take on where the hours are best spent

  • If you build things: the current cloud AI developer or engineer exam for whichever platform your work already runs on. Their value is less about the models and more about deploying and monitoring them, which is where projects tend to fall over.
  • If you need AI literacy for work: a short, practical fundamentals course is a reasonable way to meet an Article 4 expectation. On its own it won't carry much weight in hiring.
  • For everyone: a public write-up or repository of something you built, what broke and how you fixed it still says more than a certificate. Certify in the gap that project shows you.
Where to spend the hours, by what you need.
What do you need? Building AI systems: a current cloud AI developer exam. AI literacy for work: a short practical course. Credibility: a public write-up of a real build.

Scroll sideways to see the whole diagram.

Figure 2. Where to spend the hours, by what you need.

Exam names, dates and formats change often. Check the provider's own certification page before booking anything.

Where X-GenAI fits

Our AI education work includes practical AI literacy sessions for teams, built around the tools they actually use.

Sources

AWS Certified Machine Learning, Specialty, AWS certification page (retirement notice).

Microsoft Learn: retired certification exams.

Regulation (EU) 2024/1689 (Artificial Intelligence Act), Article 4.

Get new posts by email

AI updates, trend breakdowns, and picks for the certifications worth your time, a few times a month, when something's worth sending.

Have something like this to automate?