Learn about AI
Dive deep into the world of Artificial Intelligence in eLearning with our curated collection of publications and blog posts. From understanding the basics of AI to exploring its transformative impact on eLearning, our resources provide valuable insights for educators, learners, and industry professionals alike.
Featured Publications
Model Update – July 2026
AIReady has always had one goal: to help learning teams seamlessly integrate AI into their eLearning experiences without adding complexity for designers or learners.
To date, GPT-4o has provided a strong foundation for our platform. However, our priority isn’t chasing the largest models—it’s choosing the right tool for the job. We upgraded to GPT-5.4 Mini after benchmarking it against the structured, high-volume interactions AIReady actually runs, where it clearly proved to be the better fit.
Users still retain the flexibility to manually change or hardcode the model used in their interactions. This is not recommended, however, as our team balances new advancements with rigorous, use-case-based benchmarking to ensure the default model is always the best fit for your courses.
What to Expect in Your Current Implementations
As we transition our underlying technology, you may wonder how this impacts your live/future courses. Because GPT-5.4 Mini is optimized for quick, back-and-forth learning moments—like coaching prompts, feedback activities, and roleplay simulations—you can expect the following immediate changes:
- Enhanced Interaction Speed: Learners will notice a faster, more fluid, and natural conversational cadence due to reduced latency.
- Increased Behavioral Consistency: While your AI coach’s core persona and feedback style remain constant, the new model offers superior, highly predictable adherence to your custom rubrics, guardrails, and specific formatting requirements.
- Do I need to retest my interactions? This transition is designed to be seamless and shouldn’t impact your current implementation. However, we suggest performing a quick spot-check on your high-stakes or complex interactions to ensure these technical nuances align perfectly with your learning objectives.
GPT-4o helped us build a dependable platform. GPT-5.4 Mini helps us make it faster, smarter, and easier for L&D teams to implement at scale.
Questions? Reach out to us at aiready@arthalearning.com
A Guided Pathway to Responsible AI Interactions in Learning
For the past two years, learning leaders have debated whether generative AI belongs in their design workflows at all. That debate is now changing.
Accrediting bodies — the institutions that determine whose education actually counts — are beginning to publish formal guidance on how AI can and cannot be used in credentialed programs. The Accreditation Council for Continuing Medical Education (ACCME) recently released guidance for AI in continuing medical education. Similar conversations are underway in finance, law, engineering, and corporate compliance.
Whatever industry you operate in, this shift matters. It signals that AI in learning has graduated from experimental curiosity to accountable practice.
The question for every learning organization is no longer whether to integrate AI. It is how to do so in a way that stands up to scrutiny: from auditors, learners, and your own future standards.
Responsibility Lives in a Stack, Not a Platform
The most useful reframe I can offer is this: responsible AI in learning is not a feature anyone can sell you. It is a stack of decisions, some made by the technology provider and others made by your organization.
Many public conversations blur those responsibilities. That is why conversations about “AI safety” often end in either vendor-shopping or paralysis.
Across nearly every serious framework being published today, the same 5 layers appear. Understanding who owns each one is how defensible AI programs actually get built.
1. Source Integrity
Most reputational damage from AI in education traces back to one failure: the model invents something.
Hallucinations are not simply bugs to be patched. They are a known property of large language models when asked to answer outside trusted source material.
A strong platform constrains the model to vetted content, often through retrieval-augmented generation (RAG), so responses are grounded in approved documents.
Your responsibility is deciding what qualifies as approved in the first place. No vendor can determine which clinical guideline, policy memo, or internal procedure reflects your organization’s position.
That remains editorial judgement — and it stays with you.
2. Human Oversight
AI should support professional judgment, not replace it. Some version of that sentence appears in nearly every set of guidelines being published right now.
In practice, that means two things.
- Before deployment, qualified reviewers test the system against realistic scenarios and approve how it behaves.
- After deployment, someone monitors performance drift as models, documents, and learner behaviors evolve.
Platforms can provide tools such as logs, prompt controls, and test environments. Organizations must provide reviewers, governance cadence, and sign-off authority.
3. Data Stewardship
Personally identifiable information and confidential content do not belong in public AI systems.
This rule sounds obvious, yet it is frequently broken when busy employees paste sensitive material into free tools without considering governance implications.
A credible platform should provide:
- Closed and encrypted infrastructure — clear data residency controls
- Contractual assurance that customer data is not used to train public models.
Your organization must still define what can be uploaded, what must be scrubbed, and which tools are approved.
4. Transparency
Learners deserve to know when they are interacting with AI, what tool is being used, and whether a human has reviewed the output.
This is not only an ethical norm; it is increasingly a regulatory expectation.
Disclosure is primarily an authoring decision. The platform should make it simple to add disclosures within the interface and clearly indicate which model and version are being used. The wording itself — and how prominently it’s presented for different audiences and contexts — should be determined by the provider.
5. Governance and Continuous Improvement
This final layer is the one many organizations skip.
Governance determines:
- Who can build AI-enabled learning experiences
- Which tools are approved
- How new use cases are piloted
- How systems are reviewed over time
- What happens when something goes wrong
This layer has no vendor counterpart. No platform can govern your organization for you.
Deploying AI without governance is an avoidable risk.
The Advantage of Building Now
The organizations that will earn learner trust over the next decade are not the ones using the most AI. They are the ones that can show their work: source-controlled, reviewed, disclosed, and governed.
That posture is much easier to establish during pilot programs than after multiple unvetted tools are already in production.
Teams that succeed usually do three things well:
- They select platforms whose architecture aligns with the governance standards they’ll ultimately need to demonstrate. That means closed models, vetted data sources, and clearly documented data residency — substantiated in practice, not just in marketing. Switching vendors after an initial audit is costly; making the right choice upfront isn’t.
- They establish review and disclosure protocols before any learner-facing rollout — not after. A concise, one-page standard that specifies the review cadence, disclosure language, and escalation path is far more valuable than a policy written reactively after something goes wrong.
- They approach AI governance as an educational discipline, not just an IT requirement. The people responsible for content quality should also own AI quality because in an AI-enabled environment, those responsibilities are inseparable.
By the time your accreditor, regulator, or board asks how your AI-enabled learning meets their standards, the answer should already be documented. Guidance from organizations like ACCME offers an early glimpse of that conversation — treat it accordingly.
Pathways to Pilot Responsibly
Organizations ready to move from principle to practice can start in three practical ways:
See the principles in action:
Two short educational demos — Socratic Coaching and Knowledge-Driven Document Retrieval — illustrate what source-grounded, oversight-ready AI interactions look like inside a real learner experience. They’re the quickest way to make the framework tangible.
Review the technical protocols:
The AIReady security and data privacy overview outlines the encryption, data residency, model isolation, and audit posture underpinning this approach — at the level of detail your IT, legal, and compliance teams will expect. Share it directly with them; it’s written for that audience.
Talk through your use case:
If you are scoping an AI-enabled learning program — whether a coaching simulation, knowledge assistant, or dynamic assessment — and want to pressure-test it against the five-layer stack before commiting, schedule a working session with us here, review structure, disclosure approach, and the architecture required to support it.
Conclusion
Waiting may feel safer. In many cases, it is not.
AI-enabled coaching, simulations, and knowledge assistants are already among the most effective learning interventions available when built responsibly. They can shorten time-to-competency, personalize practice at scale, and surface learner insights traditional methods often miss.
The organizations standing still are not necessarily being careful. They may simply be falling behind.
This framework is not a reason to hesitate, it is a reason to move forward with confidence. If you would like help taking the first step, we would be glad to help.
References
Accreditation Council for Continuing Medical Education (ACCME). (2026). Guidance on the Responsible Use of Artificial Intelligence (AI) in Accredited Continuing Education (CE). https://accme.org/resource/guidance-on-ai/
European Parliament and Council of the European Union. (2024). Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union. https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng
National Institute of Standards and Technology (NIST). (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). U.S. Department of Commerce. https://www.nist.gov/itl/ai-risk-management-framework
AIReady. Security and Data Privacy. Artha Learning Inc. https://aiforlnd.com/security-and-data-privacy/
AIReady. Socratic Coaching — interactive demo. Artha Learning Inc. https://arthademos.s3.us-east-2.amazonaws.com/AIReady+Socratic+Coaching/story.html
AIReady. Knowledge-Driven Document Retrieval — interactive demo. Artha Learning Inc. https://arthademos.s3.us-east-2.amazonaws.com/Document+Retrieval/story.html
How to Create an AI Feedback Bot with FeedbackReady
Reading time: 10-15 minutes
In this tutorial, you’ll create a feedback bot that evaluates learner responses to the question “Explain the water cycle in your own words.” By the end, you’ll have a working bot that provides personalized, criteria-based feedback—integrated into Storyline or Rise 360.
What You’ll Need
- An AIReady API key (contact us at aiforlnd.com/contact-us if you haven’t purchased one yet)
- Articulate Storyline or Rise 360
- Your assessment question and a model answer
Creating Your Feedback Bot
Head to the CodeReady homepage (aiforlnd.com/CodeReady) and select FeedbackReady.
Enter Your API Key
Paste your AIReady API key into the field.
Enter Your Question
This is the question learners will answer. It appears in the chatbot interface, so write it exactly as you want learners to see it.
For our example: Explain the water cycle in your own words.
Be specific about what you’re asking. Instead of “Explain photosynthesis,” try “In 2-3 sentences, explain how plants convert sunlight into energy through photosynthesis.”
Provide the Model Answer
Enter what a good response looks like. The AI compares learner answers against this to evaluate quality.
For our water cycle example:
Water is always moving in a big loop. The sun heats up water in oceans, lakes, and rivers, turning it into vapor that rises into the air—that’s evaporation. As the vapor goes higher where it’s cooler, it turns back into tiny water droplets that form clouds. This is condensation. When the clouds collect enough droplets, the water falls back down as rain, snow, or hail—that’s precipitation. The water then collects in oceans, lakes, rivers, and underground, and the whole cycle starts over again.
Your model answer doesn’t have to be the only correct answer—it represents what a strong response looks like. The AI uses it as a reference point, not a rigid template.
Configure Feedback Settings
Choose how your bot delivers feedback:
Setting | Our Choice | Why |
Purpose | Corrective Assessment | We want to identify gaps and guide improvement |
Tone | Encouraging and Supportive | Keeps learners motivated while addressing weaknesses |
Response Length | Medium | Enough detail to be helpful without overwhelming |
Other tone options include “Direct and Constructive” for more straightforward feedback, or “Detailed and Analytical” for comprehensive breakdowns.
Add a Rubric (Optional but Recommended)
A rubric helps the AI provide specific, criteria-based feedback instead of general comments. Here’s an example rubric for our water cycle question:
Rubric: Explain the water cycle in your own words
Score | Level | Criteria |
4 | Excellent | Accurately describes all major stages (evaporation, condensation, precipitation, collection). Shows clear connections between stages. Uses own language demonstrating genuine understanding. |
3 | Proficient | Describes most stages correctly. Shows reasonable connections between stages. Minor gaps or imprecise language, but solid overall understanding. |
2 | Developing | Identifies some stages but missing key components. Connections between stages unclear or absent. May rely on memorized phrases without deeper understanding. |
1 | Beginning | Shows minimal understanding of the water cycle. Missing most stages. Contains significant misconceptions. |
Without a rubric, the AI provides general feedback. With a rubric, it can say things like “You scored a 3 (Proficient) because you correctly identified evaporation and precipitation, but didn’t explain condensation clearly.”
Add Example Student Answers (Optional)
You can provide sample responses at different quality levels to help the bot calibrate its feedback. This step is optional, but examples help the bot be more accurate.
For instance, you might include:
A strong response that earns a 4
A weak response that earns a 1 or 2
The feedback you’d give for each
This teaches the bot what different quality levels look like in practice.
Review the Prompt
FeedbackReady generates a prompt automatically. You can edit it to fine-tune how the AI provides feedback—for example, adding instructions like “Always suggest one specific improvement” or “Reference the rubric criteria in your feedback.”
Choose Your Output Format
Select the format that matches where you’ll use your feedback bot:
Format | Best For |
Storyline JavaScript | Articulate Storyline projects where you want full control over the interface |
Generate Code for Rise | Rise 360 courses—includes a ready-made chat interface |
Web Link | Quick testing, sharing with stakeholders, or embedding anywhere |
If you choose Storyline:
Note the variable names shown (default: TextEntry for input, GPT_Response for output). You’ll need these exact names when setting up Storyline.
If you choose Rise:
You can customize the header and button colors to match your brand.
Generate Your Code
Click generate. Your code appears—copy it and you’re ready for integration.
Adding Your Feedback Bot to Your Course
Choose your platform below.
For Articulate Storyline
Step 1: Create your variables
Open the Variables panel and create two text variables:
TextEntry (for learner input)
GPT_Response (for the AI’s feedback)
These names must match exactly what you set in FeedbackReady.
Step 2: Build the interface
On your slide, add:
A text entry field bound to the TextEntry variable (this is where learners type their answer)
A text box that displays %GPT_Response% (this shows the AI’s feedback)
A button labeled “Submit” or “Get Feedback”
Step 3: Add the trigger
Select your button and create a new trigger:
Action: Execute JavaScript
When: User clicks
Paste your generated code into the script editor.
Step 4: Test it
Preview your slide. Type an answer, click the button, and watch your feedback bot evaluate the response.
Note: If the preview doesn’t work, try publishing to Review 360 for full functionality.
For Articulate Rise 360
Step 1: Add a Code Block
In your Rise lesson, click to add a new block and select Code.
Step 2: Paste your code
Copy your generated HTML from FeedbackReady and paste it into the Code Block. The chatbot will display your question automatically, prompting learners to submit their answer.
Step 3: Lock the Continue button (optional)
Want to ensure learners actually submit an answer before moving on? You’ll need to configure two things:
On the Code Block: Click the block, open settings, and enable Set Completion Requirements
Add a Continue button: Add a Continue block below your feedback bot, then configure it to “Complete block directly above”
⚠️ Both settings are required. If you only do one, the Continue button won’t lock.
For Web Link (Quick Access)
Select the Web Link option in FeedbackReady, and you’ll receive:
A direct URL to your hosted feedback bot
An embed code for adding to other platforms
This is perfect for sharing with SMEs for review, testing before full integration, or embedding in platforms that support iframes.
Troubleshooting
The feedback bot doesn’t respond in Storyline
Check that your variable names match exactly (they’re case-sensitive)
Check that your domain is whitelisted in your AIReady account settings
The Continue button doesn’t lock in Rise
Verify “Set Completion Requirements” is enabled on the Code Block
Make sure the Continue button is configured to complete the block above
Both settings must be active
Feedback is too generic
Add a detailed rubric with specific criteria
Include example student answers at different quality levels
Make your model answer more comprehensive
Feedback doesn’t match expected quality levels
Review your rubric—are the criteria clear and distinct?
Add more example answers to help calibrate the bot
Edit the prompt to emphasize specific evaluation criteria
Ready to build?
Visit the CodeReady homepage at aiforlnd.com/CodeReady to create your feedback bot.
Want AI coaching chatbots instead?
Check out our CoachReady tutorial to create bots that answer questions and guide learners.
How to Create an AI Coach with CoachReady
Reading Time: 5-10 minutes
In this tutorial, you’ll create “Learning Buddy”—an AI coaching chatbot that helps learners understand workplace safety policies. By the end, you’ll have a working chatbot integrated into Storyline or Rise 360.
What You’ll Need
- An AIReady API key (contact us at aiforlnd.com/contact-us if you haven’t purchased one yet)
- Articulate Storyline or Rise 360
Creating Your AI Coach
Head to the CodeReady homepage (aiforlnd.com/CodeReady) and select CoachReady.
Enter Your API Key
Paste your AIReady API key into the field. This powers the AI behind your chatbot.
Configure Your Coach’s Personality
This is where your coach comes to life. For our safety policy example:
Setting | Our Choice | Why |
Name | Learning Buddy | Friendly and approachable |
Purpose | Learning Assistant | We want it to explain and answer questions |
Personality | Friendly and Conversational | Matches our informal training style |
Response Length | Medium | Detailed enough to be helpful, short enough to read quickly |
Your choices here will vary based on your needs. Building a technical support bot? Choose “Technical Support” and “Stoic and Concise.” Creating a motivational coach for a wellness course? Choose “Motivational Coach” and “Friendly and Conversational.”
Add Context (Optional but Recommended)
This is the secret to a great coach—giving it specific knowledge about your topic.
You have two options:
Option A: Paste text directly.
Copy in the key information you want your coach to know. For our safety policy coach, we might paste the company’s safety guidelines, common procedures, or FAQ answers.
Option B: Select an uploaded document.
If you’ve already uploaded documents to the AIReady portal (aiforlnd.com/portal), you can select them here.
Without context, your coach relies on general AI knowledge. With context, it becomes an expert on your specific content.
Review the Prompt
CoachReady generates a prompt automatically—this is the instruction that tells the AI how to behave.
You can edit this if needed, but for most cases the auto-generated prompt works well. You might tweak it to add specific instructions like “Always remind users to consult their supervisor for site-specific questions” or “End each response with a safety tip.”
Choose Your Output Format
Select the format that matches where you’ll use your coach:
| Format | Best For |
| Storyline JavaScript | Articulate Storyline projects where you want full control over the interface |
| Generate Code for Rise | Rise 360 courses—includes a ready-made chat interface |
| WebLink | Quick testing, sharing with stakeholders, or embedding anywhere |
If you choose Storyline:
Note the variable names shown (default: TextEntry for input, GPT_Response for output). You’ll need these exact names when setting up Storyline.
If you choose Rise:
You can customize the header and button colors to match your brand.
Generate Your Code
Click generate. Your code appears—copy it and you’re ready for integration.
Adding Your Coach to Your Course
Choose your platform below.
For Articulate Storyline
Step 1: Create your variables
Open the Variables panel and create two text variables:
TextEntry (for learner input)
GPT_Response (for the AI’s response)
These names must match exactly what you set in CoachReady.
Step 2: Build the interface
On your slide, add:
A text entry field bound to the TextEntry variable
A text box that displays %GPT_Response% (this shows the AI’s answer)
A button labeled “Ask” or “Send”
Step 3: Add the trigger
Select your button and create a new trigger:
Action: Execute JavaScript
When: User clicks
Paste your generated code into the script editor.
Step 4: Test it
Preview your slide. Type a question, click the button, and watch your coach respond.
Note: If the preview doesn’t work, try publishing to Review 360 for full functionality.
For Articulate Rise 360
Step 1: Add a Code Block
In your Rise lesson, click to add a new block and select Code.
Step 2: Paste your code
Copy your generated HTML from CoachReady and paste it into the Code Block. That’s it—your chatbot is live.
Step 3: Lock the Continue button (optional)
Want to ensure learners actually use the coach before moving on? You’ll need to configure two things:
On the Code Block: Click the block, open settings, and enable Set Completion Requirements
Add a Continue button: Add a Continue block below your chatbot, then configure it to “Complete block directly above”
⚠️ Both settings are required. If you only do one, the Continue button won’t lock.
For Weblink (Quick Access)
Select the Weblink option in CoachReady, and you’ll receive:
A direct URL to your hosted chatbot
An embed code for adding to other platforms
This is perfect for sharing with stakeholders for review, quick testing before full integration, or embedding in platforms that support iframes.
Troubleshooting
The chatbot doesn’t respond in Storyline
- Check that your variable names match exactly (they’re case-sensitive)
- Check that your domain is whitelisted in your AIReady account settings
The Continue button doesn’t lock in Rise
- Verify “Set Completion Requirements” is enabled on the Code Block
- Make sure the Continue button is configured to complete the block above
- Both settings must be active
Responses are getting cut off in Storyline
- Storyline text variables have character limits
- Try setting Response Length to “Short” or “Medium” in CoachReady
Ready to build?
Visit the CodeReady homepage at aiforlnd.com/CodeReady to create your AI coach.
Want AI-powered assessment feedback?
Check out our FeedbackReady tutorial to create bots that evaluate open-ended responses.
Featured Webinars
Webinar Name
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec iaculis nisi viverra mattis porta. Mauris quis tincidunt lacus.
Webinar Name
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec iaculis nisi viverra mattis porta. Mauris quis tincidunt lacus.
Webinar Name
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec iaculis nisi viverra mattis porta. Mauris quis tincidunt lacus.