Loading...
Loading...
Browse all stories on DeepNewz
VisitEffectiveness rating of the DALL-E 3 Detection Tool in its first year
Below 90% effectiveness • 33%
90% to 98% effectiveness • 33%
Above 98% effectiveness • 33%
Published studies or reports evaluating the tool's performance
OpenAI Launches DALL-E 3 Detection Tool with 98% Accuracy, Combats Disinformation
May 7, 2024, 01:05 PM
OpenAI has launched a new tool designed to detect images created by its DALL-E 3 generator, boasting a 98% accuracy rate for unaltered images. The tool is part of OpenAI's efforts to combat disinformation, especially with concerns that AI-generated content could impact upcoming elections, and will be shared with disinformation researchers. Additionally, OpenAI has joined forces with Microsoft and Adobe in a content credentials group. In a separate development, OpenAI has inked a content licensing deal with Dotdash Meredith, one of the largest digital publishers in the U.S. This partnership includes licensing Dotdash Meredith's archive to train OpenAI's algorithms and a strategic collaboration to develop new products, including advertising tools.
View original story
Below 90% accuracy • 33%
90% to 98% accuracy • 34%
Above 98% accuracy • 33%
Below 75% accuracy • 33%
75% to 90% accuracy • 33%
Above 90% accuracy • 33%
Highly Effective • 33%
Moderately Effective • 33%
Ineffective • 34%
No integration • 33%
Partial integration • 34%
Full integration • 33%
Yes • 50%
No • 50%
Highly Effective (>75% accuracy) • 25%
Moderately Effective (50-75% accuracy) • 25%
Low Effectiveness (25-49% accuracy) • 25%
Ineffective (<25% accuracy) • 25%
Not effective • 25%
Moderately effective • 25%
Effective • 25%
Highly effective • 25%
Highly effective • 33%
Somewhat effective • 33%
Not effective • 33%
Significantly more effective than existing tools • 33%
About the same effectiveness as existing tools • 34%
Less effective than existing tools • 33%
Highly effective • 33%
Moderately effective • 33%
Ineffective • 33%
Highly effective • 25%
Moderately effective • 25%
Slightly effective • 25%
Not effective • 25%
Yes • 50%
No • 50%
1-2 new major tech companies • 34%
3-5 new major tech companies • 33%
More than 5 new major tech companies • 33%