AI in Moodle
Using AI Effectively in Moodle and e-Learning – Learn More Efficiently, Create Content More Intelligently, and Tailor Learning Processes to Individual Needs.
Here you will find the following topics
What can Moodle do with AI right out of the box?
Starting with Moodle 4.5, four AI features are available directly in the core without the need to install a plugin. They are managed through the AI subsystem and appear in two locations: the text editor and the " Course Support" section. You can use permissions to determine which features are available to which user groups.
Generate text in Moodle
After activation, an AI icon—the now-familiar “Sparkle” icon—appears in the text editor (TinyMCE). Clicking it opens a prompt window where trainers—or, depending on their permissions, participants as well—can describe the text they need. Within a few seconds, the connected AI model provides a suggestion, which is inserted into the editor with another click.
Typical uses: Introductions to course pages or lessons, assignment prompts, discussion questions for forums, quiz questions, explanatory texts for complex concepts, and summaries for weekly or module overviews.
Generating images in Moodle
Using the same text editor, Moodle generates images on demand—typically via OpenAI’s DALL·E 3 or a comparable image model from the connected provider. You can choose the quality (Standard or High) and format (Square, Landscape, Portrait). Alt text is generated automatically, which improves the accessibility of course content (WCAG 2.2 AA).
Typical uses: Course thumbnails, header images for course sections, illustrations for abstract concepts, infographic designs for learning materials.
Summarizing content in Moodle (Summarize)
Using the “Course Support” feature, learners can click a “Summarize” button on any course page—whether it’s a text page, quiz, forum, or lesson. The AI condenses the page’s content into a concise overview. This is especially helpful for long texts, dense technical vocabulary, or before exams.
Typical uses: Quick review before exams, getting an overview of extensive onboarding modules, preparing for live sessions, and getting started with new topics.
Explaining content in Moodle (Explain, since Moodle 5.0)
The "Explain" feature provides a clearer, alternative way of phrasing the page's content—tailored to learners with varying levels of prior knowledge, non-native speakers, or when the material is particularly challenging. The "Explain" feature is not a translation, but rather an educational rephrasing.
Typical applications: Inclusive education, heterogeneous learning groups, multilingual workforces, vocational schools, and adult education.
Moodle doesn't have its own AI. But it does have an architecture that guarantees your digital sovereignty.
Moodle's AI Principles
Other LMS platforms sell you a black-box AI that locks you into a specific provider. Moodle takes the opposite approach: Since version 4.5, the core has included a complete AI subsystem—no plugin, no add-on—that you can customize according to your own rules. Four features make all the difference
Granular control
You decide where and for whom AI is available. And you determine which Moodle instances should be used for what.
Freedom of Choice of Provider
You can integrate OpenAI, Azure AI, DeepSeek, or a local model—and switch providers as the market changes. No vendor lock-in, no migration required for learners.
Compliance by Design
By default, AI is disabled in Moodle. The AI usage policy is displayed directly at the "point of use." AI reports for the GDPR and the EU AI Act are built in.
Full transparency
For Moodle Core functions, the following applies: Every AI call goes through the central AI service, which generates audits. You can see who processed what, when, and using which model—including token usage.
How does AI work technically in Moodle?
The technical elegance behind Moodle's AI approach lies in its clean separation into four layers. Each layer has a clearly defined role, and it is precisely this separation that distinguishes a controllable platform from a black box.
Layer 1: UI layer (placements)
This is where you decide where AI features appear in the LMS—currently, these are the text editor and course support in the Core. Plugins can add additional locations: in the quiz editor, in the forum, when grading submissions, in the dashboard, or as a chat widget. Each location can be enabled or disabled independently.
Layer 2: Central AI Service
The AI Service orchestrates all AI calls within the system. It receives the request from the UI, forwards it to the appropriate provider, logs an audit entry (who, when, which model, what context, how many tokens), and returns the result. This layer is key to ensuring the traceability of every AI use in Moodle.
Layer 3: Core Components
Caching, reporting, standardization of data formats, fallback logic, rate limiting, and action classes (generate_text, generate_image, summarise_text, explain_text) — the invisible infrastructure that keeps the interaction between ad placements and providers running smoothly.
Layer 4: Provider Plugins
The actual connections to external or internal AI services:
- OpenAI — Direct Access to GPT Models and DALL·E 3
- Azure AI — OpenAI models in Microsoft data centers (often with EU data residency)
- DeepSeek — available since Moodle 5.1, a powerful open-weight model
- Self-hosted via OpenAI-compatible endpoints — Ollama, LiteLLM, and LocalAI enable users to run open-source models (Llama, Mistral, Qwen, DeepSeek, and others) on their own hardware and offer Moodle as an OpenAI provider
- Specialized cloud services such as Groq for particularly fast inference
Providers specify which actions they support. The system automatically selects the appropriate provider for each action—so you can use a powerful cloud-based model for “Generate Text” and a cost-effective local model for “Summarize” without users even noticing.
Our eLeDia.ai Developments for Moodle
eLeDia.ai | Tutor
With the AI-powered tutor, your users can interact directly with the learning content. This promotes effective learning and simplifies access to complex content.
eLeDia.ai | Translate Plugin
With the AI-powered translation features, you can easily translate the learning content in your Moodle installation into various languages.
eLeDia.ai | Adaptive Learning
By integrating AI, adaptive learning paths can be created directly within Moodle. This enables personalized learning—and significantly boosts motivation to learn.
Blog Posts | AI in eLearning and Moodle
Artificial intelligence (AI) is fundamentally transforming digital learning—from the automatic generation of learning content to personalized learning paths and intelligent assistance features. On this page, you’ll find a growing collection of articles, real-world examples, and information about AI in e-learning, as well as specific ways to use AI in Moodle. Discover how AI supports educators, makes learning processes more efficient, and opens up new possibilities for digital learning.
AI in Moodle's Core: An Overview of Features, Architecture, and Full Control
Since Moodle 4.5, AI has been integrated directly into the core. Learn about the features “AI Moodle” offers, how the AI subsystem is structured, and how institutions can maintain full control over its deployment, data protection, and use
Data-secure AI for Moodle: eLeDia develops RAG architecture and AI tutor for Moodle
With a data-secure AI architecture and the eLeDia.AI Tutor, eLeDia combines modern RAG technology with Moodle. The system generates and explains learning content in a context-sensitive, local, and GDPR-compliant manner—for universities, companies, and educational providers,
AI and learning
AI and Learning – What Does This Mean for Moodle? Artificial intelligence (AI) has been the talk of the town for over a year—but what role can it play in education? The
Subscribe to our newsletter on AI and Moodle.
Four times a year, we send out a newsletter to keep you updated on the latest developments in AI and Moodle. You can unsubscribe at any time.
FAQ – Answers to Questions About AI in Moodle
Does Moodle have its own AI?
No, and that’s by design. Moodle doesn’t develop its own language models; instead, starting with version 4.5, it provides a complete AI subsystem in the core that you can connect to any provider. You decide which AI actually does the processing—OpenAI, Azure, DeepSeek, a local Llama model, or a combination. This separation of the platform and the AI model is the key difference from LMSs that come with a hard-coded black-box AI.
Which AI providers does Moodle support out of the box?
OpenAI, Azure AI, and, since Moodle 5.1, DeepSeek. The OpenAI-compatible interface also allows you to use Ollama, LiteLLM, LocalAI, and cloud services such as Groq.
Does our data stay in Germany?
Yes, if you'd like. We host open-source LLMs in our German data center as an OpenAI-compatible endpoint for Moodle. No data leaves the EU.
Is AI enabled by default in Moodle?
No. AI is disabled by default in Moodle Core. You must explicitly enable it for each feature, role, and course. The security of Moodle hosting is based on several coordinated measures. These include regular security updates, the use of firewalls and clearly defined access controls, as well as encrypted data transmission via HTTPS. This is supplemented by structured backup and recovery strategies to protect against data loss. Overall, professional hosting helps significantly reduce typical security risks.
How much does AI cost in Moodle?
The cost of AI depends heavily on usage. Tokens serve as the basis for calculating all AI costs. Prices per token vary depending on the AI model and, in some cases, also depend on whether they are output tokens or input tokens.
The following applies to Moodle Core functions: Every AI call goes through the central AI service, which generates audits. There, you can see who processed what, when, and with which model, including token usage.
Since which version of Moodle has AI been included in the core?
The AI subsystem was introduced with Moodle 4.5. Moodle 5.0 added the "Explain" feature and AI Reports. Moodle 5.1 introduced DeepSeek support and the ability to enable or disable AI features directly on a per-course and per-activity basis.
Is AI in Moodle GDPR-compliant?
The necessary features are built in. Moodle does not enable AI by default; before the first use, it displays an AI usage policy that must be actively accepted, and it logs every AI request in the AI Reports. Whether your specific implementation is GDPR-compliant depends on the provider you choose and the configuration. With locally hosted open-source models in a German data center, GDPR compliance can be easily achieved.
Does Moodle AI meet the requirements of the EU AI Act?
Moodle’s AI Principles are explicitly based on the EU AI Act, the OECD AI Principles, and the NIST AI Risk Management Framework. The technical requirements—transparency, audit trails, human oversight, and configurability—are built into the AI subsystem. Actual compliance depends on how you use the system; we will assist you in implementing EU AI Act-compliant solutions during a strategy workshop.
Can students use AI on their own?
Yes, if you want to—and to what extent. Using role-based permissions, you can decide whether participants are allowed to use AI features and, if so, which ones. We often see setups where learners are allowed to use “Summarize” and “Explain,” but “Generate Text” and “Generate Image” are reserved for instructors.
Can we switch AI providers later?
Yes, and that’s one of the main reasons behind Moodle’s architecture. Because the AI subsystem separates the action from the provider, you can swap out the provider without learners or instructors having to adjust their workflows. The UI stays the same; only the processing happens behind the scenes via a different endpoint.
Can AI automatically grade learners in Moodle?
We do not recommend it—and neither does Moodle. Plugins like AI Text (Catalyst) are designed to provide suggestions: the AI suggests feedback, and a human decides whether to approve, edit, or reject it. This separation between AI suggestions and human decision-making is important both ethically and under the EU AI Act for “high-risk applications” such as educational assessments.
How quickly can AI be implemented in Moodle?
If you use OpenAI or Azure as your provider, setup takes just a few hours—enter the provider, set permissions, define the policy, and you’re good to go. With a locally hosted model in our data center, it typically takes one day to a week at most, depending on the model and workload requirements. We recommend holding a strategy workshop beforehand to determine which features are appropriate for which target audience.
We tailor our solutions to your specific needs.
Do you have an idea for how you’d like to use AI? Or do you already know exactly what you need and are looking for someone to help you implement it quickly? Are you looking for a partner to brainstorm ideas with and suggest solutions?
Your trust, our responsibility
eLeDia is certified according to ISO/IEC 27001. This ensures that all areas - from hosting and individual development to support - meet the highest security standards. We guarantee the protection of sensitive data through external audits.
