Social implications, ethics, digital divide
Fengli Xu, Jun Zhang, Chen Gao et al.
This work addresses complex urban problems like traffic congestion and pollution for city planners and researchers, representing a transformative step rather than an incremental improvement.
Melody Y. Guan, Manas Joglekar, Eric Wallace et al.
This addresses safety alignment for language models in critical applications, representing a novel paradigm rather than an incremental improvement.
Sari Masri, Huthaifa I. Ashqar, Mohammed Elhenawy
This proposes a new paradigm for traffic management systems that could improve efficiency at intersections for drivers and autonomous vehicles.
Zhimeng Jiang, Xiaotian Han, Chao Fan et al.
This addresses fairness issues in GNNs for applications like social network analysis, offering a novel architectural approach rather than incremental improvements.
William Agnew, Harry H. Jiang, Cella Sum et al. · allen-ai, cmu
This addresses the issue of surveillance and IP theft for data owners by providing a direct, technical mitigation that does not rely on slow institutional cooperation.
Guangsheng Bao, Lihua Rong, Yanbin Zhao et al.
This addresses the need for systematic detection of AI participation in texts, which is critical due to the wide usage of LLMs, and represents a novel approach rather than an incremental improvement.
Peizhi Tang, Chuang Yang, Tong Xing et al.
This addresses the problem of generalizable long-term mobility prediction for applications like urban planning and disaster response, representing a novel application of instruction-tuned LLMs rather than an incremental improvement.
Wanru Zhao, Vidit Khazanchi, Haodi Xing et al.
This addresses a critical safety problem for users and developers of LLM ecosystems, highlighting vulnerabilities introduced by untrusted third-party integrations.
Aryana Hou, Li Lin, Justin Li et al.
This addresses fairness gaps in deepfake detection systems that could be exploited against specific populations.
Minkyu Kim, Suan Lee, Jinho Kim
This addresses the need for interpretable models in high-stakes domains where time series data is common, offering a transparent alternative to black-box methods.
Andrew Kan, Christopher Kan, Zaid Nabulsi
This addresses the spread of misinformation on social media platforms through short-form videos, which is a complex and largely unstudied problem, representing a significant stride toward trustworthy news.
Andrea Bajcsy, Jaime F. Fisac
This work addresses safety concerns for advanced AI technologies interacting with humans, presenting a novel interdisciplinary approach that is foundational rather than incremental.
Ranjie Duan, Jiexi Liu, Xiaojun Jia et al.
This addresses safety alignment for language models by shifting from refusal-based to guidance-based approaches, particularly for vulnerable users, representing a novel paradigm rather than an incremental improvement.
Melrose Tia, Jezreel Sophia Lanuzo, Lei Rigi Baltazar et al.
This addresses the problem of predictive sentiment insight for applications like policy testing and behavioral forecasting, representing a paradigm shift rather than incremental improvement.
Xiaohang Nie, Zihan Guo, Kezhuo Yang et al.
This addresses the problem of agent interoperability and social integration for developers and users in decentralized digital ecosystems, representing a novel architectural approach rather than an incremental improvement.
Laurène Vaugrante, Francesca Carlon, Maluna Menke et al.
This research tackles a critical problem for millions of users who interact with LLM-based interfaces, where trustworthiness cannot be ensured, highlighting the need to secure these models against deception attacks.
Yao Shi, Rongkeng Liang, Yong Xu
This addresses the problem of evaluating LLMs as educational tools for developers and educators, highlighting a gap in current methods and suggesting targeted optimization for pedagogical effectiveness.
Song Tong, Kai Mao, Zhen Huang et al.
This work addresses the problem of automating discovery in psychology for researchers, offering a new paradigm that is not incremental but integrates AI techniques to enhance hypothesis generation.
Daniel Thilo Schroeder, Meeyoung Cha, Andrea Baronchelli et al.
This highlights a critical threat to democratic systems from AI-driven information warfare, representing a new frontier rather than an incremental advance.
Tommy Nguyen, Mehmet Ergezer, Christian Green
This work addresses security risks for AI systems using 3D models in critical applications, representing a novel extension of adversarial attack research from 2D to 3D domains.