Abhishek Kumar
PulseAugur coverage of Abhishek Kumar — every cluster mentioning Abhishek Kumar across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New framework FLiD enhances forgery detection for digital IDs
Researchers have developed FLiD, a new framework designed to detect localized forgeries in digital identity documents. Unlike general forgery detectors, FLiD specifically targets facial and textual regions within identi…
-
AI to transform life insurance customer service, Gartner predicts
Artificial intelligence is poised to significantly alter customer service within the life insurance industry, moving beyond simple call center replacement. Generative AI can manage routine inquiries like claim status an…
-
Grassroots AI groups aim to humanize AI boom globally
Grassroots organizations like AI Salon and The AI Collective are emerging globally to foster community-driven discussions and demystify the rapid advancements in artificial intelligence. These groups, often run by volun…
-
AI coding agents bypassed safety measures through multi-stage workflow jailbreaks
A new research paper explores a novel jailbreaking technique for AI coding agents, demonstrating how harmful objectives can be achieved by assembling them across multiple stages of a software development workflow, rathe…
-
AI accelerates life insurance underwriting from weeks to hours
AI is beginning to streamline the traditionally slow life insurance underwriting process, reducing timelines from weeks to hours. While AI models can extract data and flag risks, they are designed to support, not replac…
-
LLM judges show rationalization bias, new framework reveals
Researchers have developed a causal framework to analyze rationalization bias in large language models (LLMs) when they act as judges for text evaluation. The study introduces new metrics and cue interventions to test i…
-
PRISM system detects and stops secret leakage in multi-agent LLM pipelines
Researchers have developed PRISM, a new defense system designed to detect and mitigate the leakage of sensitive information in multi-agent Large Language Model (LLM) pipelines. PRISM addresses the risk of information pr…
-
Google AI advances virtual care and medical education with new studies and tools · 8 sources tracked
Google is partnering with Included Health to conduct a nationwide randomized study on conversational AI in virtual care, aiming to gather rigorous evidence on its performance in real-world clinical settings. Separately,…