arXiv AI By Jacy Reese Anthis, Ken Burke, David Madras, Crystal Qian, Meredith Ringel Morris, Logan Cross, Caroline Pantofaru, Utsav Prabhu

Scaling public input on AI policy with pluralistic agents

Read the original on arXiv AI →

The paper introduces pluralistic agents—AI representatives that deliberate on behalf of individuals—to scale public input on AI policy. Experiments with 746 participants show that these agents can reliably shift policies toward their represented personas’ preferences and are viewed as useful and fair, outperforming a single AI reasoning model. The study highlights how agent influence varies with instructions and design, suggesting a cautious but promising role for pluralistic agents in expanding human participation in AI safety policy development.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv AI
Oct 21

Humans Correct Their Judgments Through Debate, but Weaker AI Models Do Not

The paper examines whether AI‑driven debate can help humans and weaker language models (LLMs) correct their judgments. Experiments show that debate improves human accuracy over prior beliefs, especially with weighted voting, but offers little benefit—and sometimes harms—accuracy for weaker LLM supervisors. The study highlights debate as a promising tool for scalable human oversight while questioning its effectiveness for fully automated AI control.

By Maria Victoria Carro, Denise Alejandra Mester, Facundo Nieto, Oscar Agustin Stanchi, Guido Ernesto Bergman, Giovanni Franco Gabriel Marrafini, Trinidad Borrell, Guido Freire, Joaquin Salvador Machulsky, Mario Alejandro Leiva, Ana Lopez, Nicolas Spinelli, Luca Nicolas Forziati Gangi, Eitan Sprejer, Federico Barrera Lemarchand, Gerardo Simari, Maria Vanina Martinez
arXiv AI
2d ago

Understanding the Effects of Multi-Agent Deliberation for Human-AI Decision-Making

The paper investigates whether multi-agent deliberation (MAD)—where several large language models (LLMs) engage in structured, multi-round discussion—helps humans make better decisions. Two user studies compare MAD to single-round AI assistance and examine the impact of adding a second AI and of multi-round interaction, finding that repeated interaction is the main driver of improved performance. The authors also explore design choices such as AI persona and human engagement level, noting that a devil’s advocate persona reduces accuracy and experience, while active human participation can boost accuracy at the cost of frustration.

By Harry Yizhou Tian, Ming Yin
arXiv AI
Sep 23

Biased AI improves human performance but reduces perceived helpfulness

The study tests deliberately biased AI assistants and finds that such bias improves human performance on tasks like misinformation evaluation, financial investment, and graduate education compared to neutral AI. However, participants undervalue biased AI and overvalue neutral AI, even when performance is similar. When two AI biases flank a participant’s perspective, performance gains are maintained while reducing the perceived cost and one‑sided influence.

By Shiyang Lai, Jiwoong Choi, Junsol Kim, Nadav Kunievsky, Yujin Potter, James Evans