Note that the first version of this classification approach was released in 2012. For more on that, you can read here.
Level |
Feature Capabilities |
Permit Requirement |
AI Life Cycle Core Principle Compliance Level |
A |
Responds to queries, prompts, and commands; references internal datasets. |
None |
Low/Elevated |
B |
Responds to queries, prompts, and commands; references internal and external datasets. |
Optional |
Elevated |
C |
Responds to queries, prompts, and commands; references internal and external datasets; minimal human intervention (human-in-the-loop). |
Optional/Mandatory |
Elevated/High |
D |
Responds to queries, prompts, and commands; references internal and external datasets; robust, long-term autonomy; self replicates; recursively self-improves; imitates human action; exceeds human intelligence. |
Mandatory |
High |
Notes
- External data sources include the internet and data available elsewhere, such as from IoT devices.