TheEduAssist e-learning design agency logo
Trust Centre

Responsible Human Oversight for AI-Supported Learning Work

Responsible Human Oversight for AI-Supported Learning Work.

This page explains how AI may support approved learning workflows while human accountability, source verification and project-specific controls remain necessary.

Discuss AI Requirements
Home / Trust Centre / Responsible AI

Responsible AI Principles

PURPOSE DEFINITION

Define why AI is being considered and what output is required.

APPROVED INPUTS

Use approved source material and permitted information.

MINIMUM NECESSARY DATA

Avoid unnecessary or sensitive information.

HUMAN ACCOUNTABILITY

Human reviewers remain responsible for approval.

SOURCE VERIFICATION

Check claims against approved reliable sources.

OUTPUT REVIEW

Review accuracy, tone, bias, cultural suitability and learning alignment.

TRANSPARENCY

Communicate AI involvement where appropriate to the engagement.

SECURITY AND PRIVACY

Use tools and workflows appropriate to the information involved.

ONGOING REVIEW

Update workflows as tools, policies and risks change.

Possible AI Use Cases

Where agreed and appropriate, AI may assist with specific workflows. Note that not every use case is applied to every project.

  • source-content organization
  • summarization of approved material
  • early outlines
  • draft learning objectives
  • question drafting
  • scenario variations
  • metadata
  • content tagging
  • translation workflow assistance
  • production assistance
  • maintenance support

Restrictions and Limitations

  • • Accuracy: AI outputs may be inaccurate and can fabricate sources or claims. Human review is required.
  • • Data Security: Confidential data should not enter unapproved tools.
  • • Legal and Cultural: Copyright and licensing require review. Bias may be present, and cultural suitability requires verification.
  • • Tool Evolution: Tool terms and behavior may change over time.
  • • Expertise: AI does not replace subject-matter expertise.
  • • Outcomes: AI does not guarantee faster delivery or better learning outcomes.
  • • Sensitivity: Regulated or sensitive content needs additional review.

Note: This information describes our operational approach and does not constitute legal or regulatory compliance advice.

Frequently Asked Questions

Does TheEduAssist use AI on every project?
No. AI use depends on the project scope, client preferences, and the specific tasks required. We do not apply AI where it is restricted by the client or unsuited to the material.
Can confidential information be entered into AI tools?
Confidential information should only be used in AI tools that have been explicitly approved for that level of data classification within the agreed project framework.
How is AI-generated content checked?
AI outputs undergo human review for accuracy, tone, bias, and instructional alignment. Final subject-matter approval remains with the client.
Can clients restrict AI use?
Yes. Clients can specify restrictions on AI use during the project definition phase, which will be documented in the Statement of Work.
Who approves AI-assisted material?
While instructional designers provide an initial review, the client remains responsible for the final subject-matter and legal approval of all content.
Can AI replace subject-matter experts?
No. Subject-matter experts are essential for verifying accuracy, providing context, and ensuring the final learning material aligns with organizational goals.
How are copyright concerns handled?
We advocate for using approved source material and reviewing outputs for potential licensing issues. Clients are responsible for final legal and copyright decisions regarding the deployed content.
Does TheEduAssist guarantee AI accuracy?
No. AI can hallucinate or provide outdated information. Human review and subject-matter verification are critical controls in our process.
Is TheEduAssist NIST compliant?
While we monitor industry frameworks such as the NIST AI Risk Management Framework to inform our practices, we do not hold a formal certification of conformance. Reference to a framework is not the same as certification.
AI Assist

AI can help you understand the page. For confirmed scope, timeline, and pricing, talk to TheEduAssist sales team.