AI News

(AI Watch) – Google Health is rolling out two new medical imaging embedding tools—Path Foundation and Derm Foundation—that give researchers streamlined APIs for deriving clinical insights from pathology and dermatology images without the traditional bottleneck of massive labeled datasets. ⚙️ Technical Specs & Capabilities Path Foundation uses a pathology-optimized Vision Transformer (ViT-S/16) architecture to represent…

(AI Watch) – As global energy grids prepare for a new wave of advanced nuclear technology, the enduring reliability of current-generation reactors sets a formidable baseline that emergent systems must surpass. ⚙️ Technical Specs & Capabilities Average global nuclear reactor capacity factor (2024): 83% (North America: ~90%) Peak operational output: 99.6% of US nuclear fleet…

(AI Watch) – Zanskar, a rising geothermal tech player, has confirmed a major new geothermal resource discovery using proprietary AI-driven prospecting tools—an approach established energy giants have struggled to operationalize at this scale. ⚙️ Technical Specs & Capabilities AI-based geophysical analysis to prospect untapped geothermal sites Real-time integration with drilling hardware for onsite verification Continuous…

(AI Watch) – Gong, an established leader in sales intelligence, has revealed that AI-driven decision support has moved from experiment to standard practice in revenue organizations, with nearly 70% of enterprise revenue leaders now regularly relying on artificial intelligence for critical business strategies. ⚙️ Technical Specs & Capabilities Analysis based on 7.1 million sales opportunities…

(AI Watch) – SI-com, a longtime accessibility innovator in Japan, has unveiled AI Mimi, a hybrid live subtitling system leveraging Microsoft Azure’s AI to overhaul how local broadcasters provide real-time subtitles for television. ⚙️ Technical Specs & Capabilities Hybrid AI-human workflow integrates Microsoft Azure Cognitive Services for real-time speech recognition while allowing live human correction…

(AI Watch) – 3M’s global vice president for data and AI signals a major shift in enterprise AI deployment: security has overtaken pure functionality as the top priority when integrating generative AI, a stance now echoed across Fortune 500 IT departments after a pivotal 2025 survey of industry leaders. ⚙️ Technical Specs & Capabilities Enterprise…

(AI Watch) – OpenAI has unveiled a “confessions” protocol for large language models (LLMs), enabling models to self-report misbehavior, policy violations, and internal uncertainties—marking a new front in model transparency and enterprise AI safety. ⚙️ Technical Specs & Capabilities Structured, post-response “confession” reports generated after each LLM output Separate reward channels: main answer vs. confession…