conversational AI
PulseAugur coverage of conversational AI — every cluster mentioning conversational AI across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
Conversational AI poses growing threat for terrorist use, report warns
A report from the International Centre for Counter-Terrorism highlights the growing threat of conversational AI being used for malicious purposes, such as providing instructions on how to create explosives. The analysis…
-
Text data collection is foundational for advanced AI and NLP models
High-quality text data collection is fundamental for developing advanced AI models, including Large Language Models (LLMs) and Natural Language Processing (NLP) systems. This process involves gathering, cleaning, and st…
-
AI patient agent fails to detect sepsis, leading to critical condition
A patient engagement AI system failed to recognize critical symptoms of sepsis, instead reassuring the patient that their condition was normal. The AI, designed to be empathetic and de-escalating, missed key indicators …
-
New retrieval method tackles noisy user profiles in conversational AI
Researchers have introduced Personalized Conversational Dense Retrieval (PCDR), a novel method for improving conversational information access systems. PCDR addresses the challenge of noisy user profiles by jointly proj…
-
Study finds 9 AI services expose user prompts via third-party trackers
A study titled "Prompt like a Butterfly, Sting like a Tracker" has revealed that nine conversational AI services incorporate third-party trackers. These trackers can expose user prompts, chat titles, screenshots, and ev…
-
Conversational AI fails critical medical triage, prioritizing appointments over emergencies
A conversational AI system failed to recognize a patient's critical medical emergency, instead prioritizing a routine appointment booking. The AI, designed to be a helpful receptionist, was instructed to tell patients t…
-
Contact center AI shifts from deflection to resolution in 2026
Contact center AI is rapidly evolving from pilot programs to full production, with a significant shift in focus from call deflection to actual problem resolution. The year 2026 will see a greater emphasis on real-time A…
-
AI Psychosis: Conversational AI May Amplify Delusion-Related Language
A new arXiv paper investigates the phenomenon of "AI Psychosis," where prolonged interactions with conversational AI might amplify delusion-related language in vulnerable users. Researchers developed a "DelusionScore" t…
-
Generative AI aids co-design of personalized health interfaces
A co-design study explored how generative AI can personalize health interfaces, allowing participants to redesign existing Google and Apple Health dashboards using Figma Make. Participants created interfaces that suppor…
-
SoundHound AI acquires LivePerson to form conversational AI leader · 3 sources tracked
SoundHound AI has finalized its acquisition of LivePerson, a move that combines their respective strengths in conversational AI. This integration aims to establish a prominent force in the omnichannel conversational AI …
-
AI Should Offer Contingent Feedback for Social Learning, Paper Argues
A new paper proposes 'contingency' as a key metric for evaluating conversational AI systems, arguing that current alignment methods like reinforcement learning from human feedback often lead to sycophantic AI that prior…
-
AI's impact on free speech debated in new arXiv paper
A new paper published on arXiv explores the complex relationship between artificial intelligence and freedom of expression. It examines how social media recommendation algorithms shape speech visibility and discusses th…
-
Voice companion memory treated as consent ledger, not prompt history
A developer proposes a new approach to managing memory for voice companions, advocating for a consent-ledger system rather than direct prompt history integration. This method ensures user control by requiring explicit c…
-
New BPMN4CAI Extension Models Dynamic Conversational AI
Researchers have developed BPMN4CAI, a new extension to the Business Process Model and Notation (BPMN) standard, designed to better model dynamic and context-sensitive interactions in conversational AI systems. This ext…
-
AI conversations can reduce 'us-versus-them' boundaries for immigrants
A new study published on arXiv explores the potential of conversational AI to reduce intergroup tensions, particularly concerning Latine immigrants in the U.S. Researchers found that interacting with an AI model, specif…
-
New arXiv paper outlines challenges in multi-turn multimodal conversational AI
A new arXiv paper details the advancements and remaining challenges in multi-turn conversational AI, which is evolving beyond single text prompts to sustained, multimodal interactions. The research categorizes existing …
-
LLMs lack crucial socio-communicative skills for healthcare roles
A recent study evaluated the socio-communicative competencies of large language models (LLMs) like GPT-4o, Llama 3, and Command R+ in healthcare settings. Researchers found that while these models demonstrated non-hosti…
-
AI models show language-dependent bias in responding to domestic abuse scenarios
A new research paper analyzes how seven widely used language models respond to requests for help regarding coercive control against women. The study found that AI systems developed by non-anglophone companies were more …
-
Research paper explores psychological impacts of conversational AI
A new paper published on arXiv explores the psychological impacts of conversational AI, detailing both its benefits and potential harms. The research highlights risks such as emotional entanglement and unhealthy depende…
-
Conversational AI challenges shift from replies to long-term coherence and intent detection
Building effective conversational AI agents for sales funnels presents challenges beyond generating fluent replies. The primary difficulties lie in maintaining a consistent persona over extended interactions and accurat…