AI Autonomy Levels
Tags: Frameworks
TL;DR
- Four categories clarify human involvement and learning: Assisted, Augmented, Automated, Autonomous.
- Helps select governance and oversight based on human-in-the-loop vs out-of-the-loop.
Why it matters for HK marketers: Matching AI type to oversight reduces risk in customer journeys and content automation.
The four categories
- Assisted Intelligence: Assists human decisions; hard-wired and not learning from interactions; human makes final decisions. Example: GPS navigation adjusting to road conditions.
- Augmented Intelligence: Learns from interactions with humans and environment; supports better human decisions; not fully autonomous.
- Automated Intelligence: Automates manual/cognitive tasks (routine or not); no human-in-the-loop; automates existing tasks. Example: Robotic Process Automation in back-office.
- Autonomous Intelligence: Learns and acts without human intervention; adapts to different situations. Examples: autonomous vehicles, robots, chatbots, virtual assistants.
Oversight implications
- Increase human oversight and testing as you move from Assisted/Augmented to Automated/Autonomous, especially in customer-impacting use cases.
4 AI categories distinguish human role and system adaptability.
So what for marketers
Classify each AI use case and set appropriate review, testing, and escalation paths before deployment in campaigns or customer touchpoints.
← Back to Knowledge Base