Multi-agent training aims to improve coordination on complex tasks

Investors have invested unprecedented amounts in artificial intelligence (AI), driven by its potential to transform various industries. However, concerns are mounting regarding the sustainability of these investments and the possibility of a market bubble, as the actual returns on these investments remain uncertain.

The emergence of AI-generated music has created a significant challenge in distinguishing between human-created and machine-generated compositions. As AI technology advances, the ability to produce music that closely mimics human creativity raises concerns about authenticity and originality in the music industry.

A new dataset named Q-Real has been introduced to evaluate the realism and plausibility of AI-generated images, consisting of 3,088 images annotated for major entities and judgment questions. This initiative aims to enhance the quality assessment of generative models, moving beyond the limitations of existing datasets that provide only a single quality score.

AI-driven personality assessments are increasingly utilized in hiring processes, leveraging affective computing to predict traits based on the Big Five model. However, ethical concerns arise regarding bias amplification from training data, which can lead to discriminatory outcomes based on attributes like gender, ethnicity, and age. A counterfactual-based framework is introduced to evaluate and quantify bias in these assessments using generative adversarial networks (GANs).

A new study introduces the Unified Attention-Mamba (UAM) backbone, designed specifically for cell-level classification of tumor cells using radiomics features. This innovative approach enhances the diagnostic accuracy of hematoxylin and eosin (H&E) images by focusing on micro-level morphological and intensity patterns, which are crucial for precise tumor identification.

The rapid growth of artificial intelligence (AI) has led to significant investments, with companies now valued at trillions of dollars. This surge is driven by optimistic projections about AI's potential to enhance economic productivity, despite emerging concerns about market sustainability and potential bubbles.

A new benchmark called CritPt has revealed that leading AI models, including Gemini 3 Pro and GPT-5, are unable to perform complex physics tasks at the level expected of early-stage PhD research, indicating significant limitations in their capabilities as autonomous scientists.

A recent study reveals that a significant number of teenagers now prefer conversing with artificial intelligence (AI) over real people, highlighting a concerning trend in social interaction among youth. This shift raises alarms about the implications for mental health and interpersonal skills as reliance on AI for communication grows.