Half of surveyed breast imaging specialists use FDA-approved AI detection tools, but the technology delivers results below expectations across all measured metrics.
A survey of 215 Society of Breast Imaging members reveals a significant gap between anticipated and actual performance of AI breast cancer detection systems. Only 35 percent of radiologists reported lower recall rates—the key efficiency metric—compared to the 59 percent who expected this improvement.
The shortfall extends beyond recall rates. Radiologists measured performance across multiple categories, and AI tools underdelivered in each one. Despite FDA approval and deployment in clinical settings, the systems have not matched the improvements radiologists anticipated when adopting the technology.
The findings suggest that while AI tools for cancer detection have gained regulatory clearance and clinical adoption, real-world performance remains a work in progress. The results raise questions about the effectiveness of current AI implementations in diagnostic imaging and the gap between laboratory validation and clinical practice.
Booksellers report that artificial intelligence companies are bulk purchasing rare books, then removing them from circulation. The practice raises concerns about content acquisition methods used to train AI models.
SpaceX's AI division has released Grok Bot, an always-on AI agent service that can independently complete workplace tasks by signing into your existing apps and tools.
Guitar manufacturer D'Addario has acknowledged using AI-generated music from Suno in a promotional video, after weeks of denial. The company had previously attributed the audio artifacts to export issues and production tools.