Search
Research Revelation: Subjecting AI to a Deluge of Low-Quality Social Media Posts May Result in Permanent Cognitive Impairment
2025-10-23 /
Read about 0 minute
Author：小编
Recently, a collaborative research team hailing from Texas A&M University, alongside other esteemed institutions, has uncovered a startling finding: subjecting large language models (LLMs) to prolonged exposure to the dregs of social media content (think hastily-crafted viral posts, sensational clickbait comments, and the like) can inflict irreparable "cognitive impairment" on these AI systems. This impairment manifests itself as a marked deterioration in fundamental capabilities, including reasoning prowess, the ability to comprehend lengthy contexts, and adherence to ethical safety standards. Furthermore, attempts to reverse this damage through subsequent training with high-quality data prove to be largely futile.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
