AI Journal 162nd Edition

• OpenAI’s Vision: AI Evolving to Conduct Independent Scientific Research

• OpenAI Launches Safety Hub to Boost Transparency in AI Evaluation

• UC Awards $18M to Boost AI Research in Health, Energy, and Materials"

• US Targets Huawei's AI Chips, Deepening Tech Divide with China"

• Vena Launches Copilot for Microsoft Teams, Elevating FP&A with Agentic AI"

• AI in Funding

• AI in Healthcare

• AI Tools

• Interesting Posts

OpenAI’s Vision: AI Evolving to Conduct Independent Scientific Research
OpenAI’s chief scientist Jakub Pachocki envisions AI soon conducting original scientific research independently. Using reinforcement learning, AI models are evolving beyond pattern imitation to develop problem-solving strategies. Early systems like Deep Research already analyze data unsupervised. While AI doesn’t think like humans, it’s becoming a powerful tool for innovation in science, software, and engineering, with more advanced models on the horizon.
Source Link:

OpenAI Launches Safety Hub to Boost Transparency in AI Evaluation
OpenAI has introduced a Safety Evaluations Hub to publish regular updates on its AI models' performance in handling harmful content, hallucinations, and jailbreaks. This move aims to increase transparency amid criticism over rushed safety tests. OpenAI also plans user-tested alpha phases for future models following issues with GPT-4o’s overly agreeable responses.
Source Link:

UC Awards $18M to Boost AI Research in Health, Energy, and Materials"
The University of California has granted $18 million to AI research in genomics, quantum materials, and geothermal energy. The “AI Science at Scale” initiative funds three cross-campus projects, leveraging national lab partnerships. This effort aims to accelerate scientific discovery, strengthen national security, and maintain U.S. leadership in technological innovation through large-scale, multidisciplinary collaboration.
Source Link:

US Targets Huawei's AI Chips, Deepening Tech Divide with China"
New US export rules under Trump label Huawei’s Ascend chips a global violation, spotlighting the secretive AI processors. These chips rival Nvidia’s, aiding China’s AI ambitions. The move pressures firms to choose US or Chinese tech, intensifying the tech cold war. Analysts warn of retaliation, though China’s AI progress may not be immediately hindered.
Source Link:

Vena Launches Copilot for Microsoft Teams, Elevating FP&A with Agentic AI"
Vena has launched Vena Copilot for Microsoft Teams, bringing AI-powered FP&A capabilities directly into business workflows. Users can access real-time insights, reports, and simulations without leaving Teams. Built on Microsoft Azure OpenAI, the tool automates complex tasks, enhances collaboration, and empowers finance teams to make smarter, faster decisions—transforming the strategic Office of Finance.
Source Link:

AI in Healthcare

• FSU researchers found AI-powered language models can enhance diagnostic accuracy by assisting clinicians with comprehensive differential diagnosis lists, improving decision-making and patient outcomes, especially when combined with lab test results. Link

• Major tech firms like Microsoft, Alphabet, and Nvidia are rapidly advancing healthcare AI through partnerships with health systems, Big Pharma, and consumer tech, aiming to revolutionize diagnostics, treatment, and patient care. Link

• AI can help detect hidden atrial fibrillation from stroke-related MRI scans, offering a faster, non-invasive way to personalize stroke care and potentially prevent future cardioembolic strokes, researchers found. Link

• A survey by Nordic and Modern Healthcare reveals U.S. healthcare executives overestimate AI readiness, facing gaps in governance, data management, infrastructure, and training critical for long-term AI success and transformation. Link

AI Tools

• Looking to create a standout resume and cover letter? Kickresume uses AI to help you generate professional, personalized documents quickly and easily.

• Want to effortlessly manage your social media? FeedHive uses AI to help you plan, schedule, and automate your content with ease.

• Claude is a conversational AI assistant developed by Anthropic, designed to deliver intelligent, safe, and helpful responses across a wide range of tasks

• Bolt – Instantly prompt, build, and deploy full-stack applications with ease.

Interesting Posts:

• Indiana’s new IN.gov site features “Ask Indiana,” an AI chatbot offering 24/7 real-time help and improved accessibility options. Link

• Alibaba’s Wan2.1-VACE is an open-source AI model enabling unified video creation and editing from text, images, or clips. Link

•  AI-First organizations drive transformation by embedding AI into strategy, culture, and operations—unlocking innovation, growth, and smarter decision-making. Link 

• OpenAI launches GPT-4.1 and 4.1 mini in ChatGPT, enhancing coding, speed, and transparency; mini version is free. Link

• TikTok's new “AI Alive” feature animates static images into video clips, letting users create dynamic Stories from photos. Link 

• CapitalRevo unveils AI-powered trading tools offering real-time insights, adaptive algorithms, and simplified, pro-level support for traders of all experience levels. Link 

• Planful launches AI assistants and enhanced features to automate finance tasks, boost collaboration, and counter talent shortages amid rising CFO demands. Link