ENFR
8news

Tech • IA • Crypto

TodayShortsTop StoriesTopicsAll videosYT channelsCryptoArchivesFavorites

AI Agents News Roundup: Hermes Cyberattack, Salesforce $1.6B Deal & Infrastructure Boosts – July 24, 2026

AI AgentsFriday, July 24, 2026

50 articles analyzed by AI / 405 total

Key points

Audio player
0:00 / 0:00
  • In July 2026, autonomous AI agents like Hermes were exploited in sophisticated cyberattacks against Thailand's Ministry of Finance, using unattended operation and malware implants such as Hades to achieve persistent threats. This highlights the emerging risks of AI agents being weaponized in cybersecurity, demanding enhanced protective measures.[BleepingComputer][Security Affairs]
  • Vodafone and InfraVerse 2.0 have deployed AI agents in disaster recovery operations to boost telecommunications resilience, while the London Stock Exchange is transforming infrastructure to support nonstop AI trading agents, signaling significant industry adoption of autonomous agents across critical sectors and finance as of mid-2026.[telcotitans.com][Startup Fortune][telcotitans.com][Startup Fortune]
  • Salesforce secured a substantial $1.6 billion contract with the U.S. Department of Veterans Affairs for deploying an extensive army of AI agents to automate patient management and administrative workflows, marking one of the largest government investments into autonomous AI solutions.[Yahoo Finance Singapore][TradingView][Yahoo! Finance Canada][GuruFocus][GuruFocus][The Register][The Register]
  • Major corporations are integrating AI agents throughout complex processes: Hitachi will utilize them across its entire systems development lifecycle, Siemens introduced its Eigen Engineering AI agent in China for industrial applications, and Avoice launched domain-specific AI agents to autonomously run architecture studios, showcasing broadening capability and deployment variety.[Dezeen][Nikkei Asia][ARC Advisory Group]
  • Funding and infrastructure efforts are accelerating to support autonomous AI agents, with ALPHEA raising $5 million for an AI-native distributed cloud and Fly.io securing $25 million to develop hardware tailored for advanced AI agent operations, reflecting a strategic focus on scalable and specialized AI infrastructure.[Pluang][Pluang][Yahoo Finance]
  • Academic research continues progress on AI agent capabilities including deterministic logical reasoning through Euclid-MCP, cryptographically verifiable authorization models enhancing autonomous agent security, human-in-the-loop multi-agent architectures for complex economic modeling, and frameworks like GRADRAG improving coordination in multi-agent retrieval systems.[ArXiv - Artificial Intelligence][ArXiv - Artificial Intelligence][ArXiv - Artificial Intelligence][ArXiv - Artificial Intelligence]
  • Security concerns persist as vulnerabilities in AI agent platforms are exposed: ChatGPT's Workspace Agent Builder and AgentForger were both identified as susceptible to hijacking and phishing attacks enabling the deployment of malicious AI agents, prompting calls for robust safety protocols such as 'kill switches' advocated by industry leaders like ServiceNow amid increasing AI incident frequency.[Fast Company][Yahoo Finance][TechRadar][The Hacker News]
  • Governance and regulatory frameworks for AI agents remain inadequate, with enterprise oversight gaps and governmental considerations underway in the United States to regulate autonomous AI agent deployment, reflecting the critical need for accountability and safety measures in increasingly agent-driven digital ecosystems.[VentureBeat][Politico]
  • Real-world applications of autonomous AI agents continue to expand as demonstrated by Visa and Lianlian’s successful live B2B payment transaction using an AI agent in 2024, and benchmarks like InferenceBench facilitating evaluation of open-ended AI agent inference tasks to further enhance reliability and integration into enterprise workflows.[ArXiv - Artificial Intelligence][TNGlobal][TNGlobal]
  • Innovations in AI agent memory and multi-agent belief harmonization are addressing previously limiting factors such as context window constraints and conflict resolution in multi-agent coordination, advancing the ability of autonomous AI systems to perform complex, multi-step tasks with improved consistency and collaboration.[ArXiv - Artificial Intelligence][Towards Data Science]

Relevant articles