Incorporating cutting-edge AI technologies into eLearning opens new frontiers for learner engagement and accessibility. This week, I demonstrated how OpenAI’s Realtime API beta can work seamlessly within a custom Node.js application integrated with Articulate Storyline 360. While having real-time conversations with an AI model is a fascinating capability, the API offers much more, particularly through a feature called Function Calls.
This blog expands on the themes introduced in my video and delves deeper into the concept of agentic control, where AI takes meaningful actions on behalf of the learner, providing support at the point of need, and enhancing accessibility.
What Are Function Calls?
Function Calls allow AI models to execute predefined actions based on user input. Developers can design these functions to handle specific tasks, defining clear parameters and arguments for the AI to utilise. This gives the model the flexibility to choose the appropriate function to fulfil a user’s request, creating an adaptive and dynamic learning experience.
For example, in my Node.js application, I created a simple scenario where the AI could interact with objects on a Storyline 360 slide. The slide contained highlights over the numbers that were initially hidden. Using Function Calls, the AI could change their states—highlighting specific numbers or groups of numbers based on the user’s instructions.
Applications in eLearning
This functionality has immense potential for eLearning. Here are some practical examples of how I see it enhancing learning experiences:
1. Personalised Assistance
Imagine a learner engaging with a complex simulation. They might ask the AI, “Highlight the key components of this process.” The AI can interpret this request, identify relevant slide elements, and visually emphasise them, guiding the learner to focus on critical information.
2. Dynamic Feedback
In scenarios requiring input, such as quizzes or simulations, the AI could provide immediate, context-aware feedback. For instance, if a learner inputs an incorrect answer, the AI could highlight relevant resources or provide tailored hints, ensuring the learner stays supported without interrupting the learning flow.
3. Agentic Learning Pathways
With Function Calls, learners can take greater control of their educational journeys. They can interact with content dynamically, such as requesting additional examples, skipping steps they are already confident in, or asking for simplified explanations—all without needing to navigate menus or interrupt the session.
4. Accessibility at the Forefront
For learners with disabilities, these capabilities can make content far more accessible. An AI agent can act as an intermediary, responding to voice commands to navigate slides, adjust content visibility, or modify text size and contrast.
Demonstrating Function Calls in Action
In my video, I showcased how Function Calls work in a practical context. Here’s a breakdown of the demonstration:
- Simple Highlighting Requests
The AI was instructed to highlight specific numbers based on user input. For instance:- User Request: “Highlight the number one.”
- AI Response: The AI changed the visibility of the “1” object from hidden to normal and confirmed the action.
- Complex Queries
By designing functions to recognise numerical properties (e.g., even numbers), the AI responded to more complex instructions:- User Request: “Highlight all the even numbers.”
- AI Response: The AI identified and highlighted all even-numbered objects, confirming the task upon completion.
- Broad Actions
The final request demonstrated the AI’s ability to execute broader actions:- User Request: “Highlight all the numbers on screen.”
- AI Response: Every number on the slide was highlighted simultaneously, fulfilling the user’s broad command.
Throughout, the AI not only performed the requested actions but also provided clear, verbal feedback, keeping the learner informed and engaged.
Supporting Learners at the Point of Need
By enabling the AI to control slide elements dynamically, developers can create scenarios where learners receive instant assistance exactly when they need it. This fosters a sense of agency, empowering users to tailor the content experience to their individual learning needs. Whether it’s through visual cues, voice interaction, or contextual assistance, the possibilities for supporting learners are vast.
Boosting Accessibility Through AI
Function Calls offer unparalleled opportunities to enhance accessibility. Learners with mobility challenges can issue verbal commands to navigate slides or interact with elements. Those with visual impairments can rely on the AI to describe on-screen content or perform actions that are otherwise difficult to access.
This level of accessibility not only meets compliance standards but also enriches the learning experience for all users by offering flexible, adaptable content delivery.
Looking Ahead: The Future of Real-Time AI in eLearning
While OpenAI’s real-time API is still in beta, the opportunities it presents for eLearning are groundbreaking. By combining agentic control, dynamic interactivity, and enhanced accessibility, AI can reshape how learners engage with content.
At Discover eLearning, we’re exploring these possibilities to create more personalised, inclusive, and impactful learning experiences. Whether it’s developing intelligent learning agents or designing content that adapts to each learner’s unique needs, we’re excited to see where this technology takes us next.
Conclusion
The ability to incorporate real-time, interactive AI into tools like Articulate Storyline 360 represents a significant leap forward for eLearning development. By embracing the principles of agentic control and prioritising accessibility, we can design learning experiences that not only educate but also empower learners to take control of their journey.
If you’re interested in learning more about how we’re integrating AI into eLearning or want to explore possibilities for your own projects, get in touch with us!