Chapter 6 - Safeguarding Students' Well-being in the Age of AI.pdf in the loop and over the loop for AI use, is there a difference?

https://drive.google.com/file/d/1l3n8YT0VvcoNG45dKyK2zUuV8UikesQT/view

https://go.gov.sg/ai-chapter6

https://www.youtube.com/watch?v=6bJRTqW0O0U

Short answer: yes — for AI discussions they usually mean different things.

“In the loop”

• Means a human is directly involved in the decision flow

• The AI cannot act without the human’s approval or input

• Focus is on control + oversight

Examples:

• Human approves every AI-generated quiz before publishing

• Doctor reviews every AI diagnosis before giving it to patient

This is often called Human-in-the-loop (HITL).

---

“Over the loop”

• Means the human is monitoring the AI system but not approving each step

• The AI acts autonomously, but the human can audit, intervene, or change settings

• Focus is on governance + supervision, not direct control

Examples:

• Teacher monitors dashboards of an adaptive learning system

• Compliance officer audits AI grading patterns monthly

• Engineer sets policies and escalation rules for LLM agents

This is sometimes referred to as Human-over-the-loop (HOTL) or oversight mode.

---

Comparison

Aspect	In the Loop	Over the Loop
Who acts?	AI acts only after human checks	AI acts autonomously
Human role	Approval / correction	Monitoring / governance
Speed	Slower	Faster
Risk	Lower	Higher (but manageable)
Use case	Safety-critical decisions	Large-scale operational systems

---

In AI policy contexts

• Academic/education AI increasingly uses in-the-loop for:
  ✔ feedback to students
  ✔ marking/assessment
  ✔ pastoral / wellbeing contexts
  because of safety + ethics

• Operational AI (recommendations, scheduling, routing) often shifts to over-the-loop when mature.

---

One more phrase

People sometimes add a third:

“On the loop”

• Like over-the-loop, but with real-time intervention

• Common in robotics, autonomous vehicles, or agentic AI

---

Bottom line

For AI use, the distinction is meaningful:

“In the loop = control”,
“Over the loop = supervision.”

If you want, I can show how these map to:

• classroom AI tools

• marking/assessment policies

• MOE/EdTech Masterplan risk levels

• AI governance maturity models

Just say which angle you want.