What's Your AI Style? Take the 2-minute quiz - are you a Cyborg, Centaur or Self-Automator? →
Manx Technology GroupSmart Island
Education · Visual

Courses Quadrant — AI Exposure × Augmentation

284 UCM courses plotted by Anthropic Economic Index scores — horizontal axis is how much of the work AI can automate, vertical is how much AI augments the human who stays. Bubble size is the live IoM vacancy count linked to the course's matched occupations. Click a bubble to open the course.

🛡️Augmented (safer)
95
AI assists, human stays central
⚠️Transitioning
44
Both automation and augmentation high — role will look different in 5 years
Distinctly Human
0
Low AI footprint either way
🔥At-Risk
145
Heavy automation, little augmentation
🛡️ Safer — Augmented
AI assists but doesn't replace
⚠️ Transitioning
High automation AND high AI leverage
✋ Distinctly Human
Low AI exposure either way
🔥 At-Risk
Heavy automation, little augmentation
Highly resilient (70+)
Resilient (50-69)
Moderate (35-49)
Higher risk (<35)
Bubble size = linked IoM vacancies
How to read this chart

Each bubble is a live UCM course. Both axes come from the Anthropic Economic Index — a task-level analysis of what AI can and can't do in each occupation:

  • Automation exposure (x-axis): percentage of tasks AI could perform on its own. Higher = more of the role's daily work is AI-performable.
  • Augmentation potential (y-axis): percentage of tasks where AI makes the human more productive. Higher = AI is a co-worker, not a replacement.
  • Bubble size: live IoM vacancies linked via the course's matched SOC 2020 occupations. Bigger bubble = more current demand on the Island.
  • Colour: the composite AI Resilience Score (blends automation, Frey-Osborne, augmentation). Green = more resilient, red = more exposed.

Filter by UCM category (apprenticeships, HE degrees, short business courses, etc) with the pills above. Hover for course details and a direct link; click any bubble to open the full course page.

Courses without Anthropic Economic Index coverage (no matched SOC, or a SOC our dataset doesn't score) are excluded from the plot to keep the signal clean.