September 11, 2025

analysis  
Anthropic Claude Enhances Documents, Nvidia Partners University

Artificial intelligence continues to reshape various industries, with Anthropic enhancing its Claude AI assistant to create and edit documents like spreadsheets and presentations, though the company cautions about potential security risks such as prompt injection attacks. Oracle is integrating OpenAI's technology into its Health Patient Portal, set to launch in 2026, to help patients understand their medical records in plain language. Walmart is also partnering with OpenAI to offer customized AI training to its U.S. employees starting in 2026. In the realm of AI security, SPLX has launched an AI Asset Management tool to map and secure AI systems, backed by $8.99 million in funding, while Jericho Security introduced an AI-powered phishing defense platform, securing $15 million in Series A funding. The University of Utah is investing $50 million in partnership with HPE and NVIDIA to advance AI research, particularly in medical fields. Salesforce CEO Marc Benioff reports strong growth for its AI agent technology, AgentForce, which has acquired over 12,000 customers and is crucial for the company's revenue. Meanwhile, the translation industry is seeing a second phase of generative AI adoption, focusing on integrated systems, according to Slator's 2025 Guide. The European Union's AI Act is facing criticism for its definition of 'general-purpose AI' and its risk-based framework. In market news, Oracle and Nvidia stocks are leading an AI-driven rally, contributing to record highs for the S&P 500 and Nasdaq.

Key Takeaways

  • Anthropic's Claude AI can now create and edit documents like spreadsheets and presentations, available to Max, Team, and Enterprise customers.
  • Anthropic warns of security risks, including prompt injection, associated with Claude's new file creation and editing capabilities.
  • Oracle will use OpenAI's models in its Health Patient Portal starting in 2026 to help patients understand medical records in simple terms.
  • Walmart is collaborating with OpenAI to provide customized AI training to its U.S. employees, launching in 2026.
  • SPLX launched AI Asset Management to secure enterprise AI systems and has raised $8.99 million.
  • Jericho Security introduced an AI phishing defense platform and secured $15 million in Series A funding.
  • The University of Utah is investing $50 million with HPE and NVIDIA to enhance AI research capabilities.
  • Salesforce's AI agent technology, AgentForce, has surpassed 12,000 customers and is a significant revenue driver.
  • The EU AI Act is facing criticism regarding its definition of 'general-purpose AI' and its risk assessment framework.
  • AI-related stocks, including Oracle and Nvidia, are driving market rallies, pushing the S&P 500 and Nasdaq to record highs.

Claude AI now creates and edits documents like spreadsheets and presentations

Anthropic's AI assistant Claude can now create and edit files, including Excel spreadsheets, Word documents, and PowerPoint presentations. This new feature allows users to perform complex analytical tasks directly within Claude.ai and the desktop app, streamlining workflows for businesses. It can generate financial models, convert invoices, and produce reports from raw notes. This capability is available to Max, Team, and Enterprise customers, with Pro users gaining access soon.

Anthropic's Claude AI adds file creation but warns of security risks

Anthropic has released a new file creation and analysis feature for its Claude AI assistant, allowing users to generate documents like spreadsheets and presentations. This tool is available to Max, Team, and Enterprise subscribers, with Pro users getting access shortly. While powerful, Anthropic warns of security risks, including potential prompt injection attacks, due to the AI's internet access within its sandbox environment. Users are advised to monitor Claude's actions closely.

Claude AI gains file editing abilities amid security concerns

Anthropic's Claude AI can now create and edit files like spreadsheets and presentations, enhancing productivity for users. However, the company has issued warnings about potential security risks, including prompt injection attacks, which could lead to data leaks. While the feature offers significant benefits for tasks like data analysis and document modification, users are advised to monitor Claude's interactions carefully and verify all generated files. Anthropic states it is refining these features through ongoing testing.

Oracle's patient portal to use AI for easier medical record understanding

Oracle is adding AI capabilities to its Health Patient Portal to help patients understand their medical records. Starting in 2026, patients will be able to get plain-language explanations of diagnoses and test results, and ask questions directly within the portal. This feature, built on OpenAI's models, aims to empower patients to manage their health more effectively. It will simplify medical terms, help prepare for doctor visits, and assist in drafting messages to clinicians, all while maintaining secure safeguards.

Oracle patient portal to feature AI for clearer medical record access

Oracle is integrating AI into its Health Patient Portal, enabling patients to understand their medical records better. The new features, planned for general availability in 2026, will allow users to ask questions about diagnoses, test results, and treatments in simple terms. Built using OpenAI's technology, the AI will explain complex medical jargon and help patients prepare for discussions with their doctors. Oracle emphasizes that patient privacy is protected and the AI will not provide diagnoses or treatment recommendations.

Walmart employees to get custom AI training via OpenAI certification

Walmart will offer its U.S. employees free access to a customized AI training program developed with OpenAI. This initiative, launching in 2026, aims to equip frontline and office-based staff with AI skills for both work and personal life. Walmart is collaborating with OpenAI to tailor the program to the retailer's needs. Employees can also access existing AI training resources through Walmart's education benefit while awaiting the new program's launch.

Slator's 2025 Guide details 15 AI uses in translation

The 2025 Slator Pro Guide highlights 15 ways Artificial Intelligence can improve translation workflows. It rates each AI application based on its potential benefit to language service providers and buyers. The guide shows the industry moving into a second phase of generative AI adoption, focusing on integrated systems rather than just experimentation. This resource provides a quick overview of AI's value and new workflow options in translation and localization.

SPLX launches AI Asset Management to secure enterprise AI systems

SPLX has introduced AI Asset Management, a new feature for its platform that helps companies map and secure their AI models, workflows, and infrastructure. This tool combines AI Bill of Materials generation, vulnerability scanning, and workflow analysis to give security and engineering teams full visibility into their AI stack. It helps identify risks, benchmark vulnerabilities, and ensures compliance with regulations like the EU AI Act. SPLX has raised $8.99 million to support its AI security solutions.

University of Utah partners with HPE and NVIDIA for AI research

The University of Utah is forming a partnership with Hewlett Packard Enterprise (HPE) and NVIDIA to boost AI research. This collaboration, involving a $50 million investment, aims to significantly increase the university's computing power for studying cancer, Alzheimer's, mental health, and genetics. The initiative seeks to create a robust AI ecosystem in Utah, driving medical breakthroughs and fostering entrepreneurship. Philanthropic gifts, including a lead donation from the Huntsman family, will support the development of AI supercomputer systems.

Salesforce CEO discusses AI growth and competition

Salesforce CEO Marc Benioff highlighted the rapid growth of his company's AI agent technology, AgentForce, which has gained over 12,000 customers and contributes significantly to revenue. Despite some investor concerns about AI demand and its impact on traditional software, Benioff remains optimistic. He also mentioned Salesforce's success in securing government contracts, competing with companies like Palantir. Salesforce aims to return to double-digit growth by effectively monetizing its AI capabilities.

Jericho Security launches AI phishing defense with $15 million funding

Jericho Security has launched an AI-powered platform to combat phishing and deepfake attacks, making advanced cybersecurity training accessible to all organizations. The platform uses generative AI to create personalized simulations across email, messaging, voice, and video, targeting sector-specific threats. The company has secured $15 million in Series A funding, led by Era Ventures, to expand its research and development. Jericho's solution is already used by the U.S. Department of Defense.

EU AI Act faces criticism over 'general-purpose AI' definition

A recent analysis suggests the European Union's AI Act may create a false sense of security due to conceptual gaps, particularly in defining 'general-purpose AI' (GPAI). The Act's terminology, like GPAI, is seen as a legal construct not organically derived from AI research, potentially misrepresenting how these systems function. Critics argue the risk-based framework struggles to quantify the unpredictable nature of AI models. The Act's approach to systemic risk also borrows from AI Safety discourse without fully engaging with its core concepts.

Nvidia and Oracle stocks lead AI rally as futures remain flat

Futures for the Dow Jones, S&P 500, and Nasdaq were little changed ahead of the CPI inflation report. However, the stock market saw a rally earlier, with the S&P 500 and Nasdaq reaching record highs. This surge was largely driven by AI-related stocks, with Oracle and Taiwan Semiconductor showing significant gains. The market experienced some weakening in the afternoon trading session.

Sources

AI Document Creation AI File Editing Anthropic Claude Spreadsheet Generation Presentation Generation Word Document Generation AI Security Risks Prompt Injection Attacks AI for Healthcare Patient Portals Medical Record Understanding Oracle Health OpenAI Models AI Training Walmart AI AI in Translation Language Service Providers Generative AI AI Asset Management AI Security AI Compliance EU AI Act AI Research HPE NVIDIA AI Supercomputing AI Ecosystem Salesforce AI AgentForce AI Competition AI Phishing Defense Deepfake Attacks Cybersecurity Training Generative AI Simulations AI Funding General-Purpose AI AI Risk Assessment AI Market Rally AI Stocks