| 摘要： | This study examined whether Microsoft Azure and human participants differed
significantly in their judgment of controversial image content (e.g., adult and
violent content). Significant differences were observed between the Azure- and
participant-awarded scores on all four indexes—Adult, Racy, Gore, and Event
Description. The artificial intelligence made stricter judgments than did the
participants. The introduction of artificial intelligence may facilitate the efficient
judgment of images by website administrators. However, the standards upheld by
artificial intelligence appear to be unique and stricter than those upheld by people.
In the short term, the boundary between art and pornography may remain difficult
to identify. Assessing the difference in judgments between artificial and human
intelligence should be a topic of continual concern. |
