Security firms are racing to develop solutions that not only track the actions of agents but also monitor the tools and plugins they incorporate in their workflows.
Google has launched the Gemini 3.8 Flash along with a specialized cybersecurity variant, Gemini 3.8 Flash Cyber. This marks their third Flash release in just six weeks. The cybersecurity edition is exclusive to trusted testers and government agencies. Compliance with the EU AI Act means that all general-purpose models must include comprehensive documentation, copyright details, training data obligations, and assessments for systemic risks.
Uber is planning to cut approximately 3,300 jobs, around 10% of its global workforce, and exit its operations in Nigeria and Uganda. The company is shifting focus and investment towards developing robotaxi technology. According to an internal memo, affected employees have been informed, except where local regulations require additional steps. In the European Union, this means consulting worker representatives and observing a 30-day waiting period after notifying authorities.
Coder has launched Agent Relay, a new self-hosted environment that enables Cursor’s cloud agents to operate directly within a customer’s infrastructure. This initiative, with SpaceXAI as the launch partner, is currently in private preview. While Cursor manages inference and planning, it remains classified as an ICT third-party provider under the EU’s Digital Operational Resilience Act.
Wonderful has successfully closed a $550 million Series C funding round, achieving a $5 billion valuation. This round was led by Insight Partners, with participation from Salesforce. The funding milestone comes eighteen months after Wonderful was founded and just six months following a $150 million raise at a $2 billion valuation. Although the press release is datelined in Amsterdam, Israeli tech media have reported Wonderful as an Israeli company.
Fireflies has launched innovative Voice Agents designed to handle screening interviews, sales discovery calls, and support conversations. These agents have facilitated over 40,000 interactions across 2,100 organizations in 97 countries. Due to their use in recruitment and candidate assessments, these systems fall under the high-risk category of Annex III in the EU AI Act, with full compliance requirements starting 2 December 2027. However, obligations related to transparency and GDPR compliance are currently in effect.
Opera contested the European Commission's decision, claiming it was wrong not to classify Microsoft as a gatekeeper for its Edge browser platform service under the Digital Markets Act (DMA). The legal challenge was unsuccessful, confirming Microsoft's position in the regulatory framework.
Until recently, Kansas' unemployment insurance system ran on software unchanged since the 1970s, forcing customer service agents to juggle multiple outdated interfaces, including one operated only via command line. The system also shut down nightly for batch processing, causing frustration for users trying to submit claims outside business hours. At the end of 2024, a modern overhaul introduced AI tools that are continuously enhancing the platform. Chatbots assist agents by quickly providing answers to complex questions unique to each claim. AI scans internal data and archives to streamline information, identify frequently asked questions for the website, and help flag fraudulent claims. Over 90% of claims now avoid agent contact, enabling staff to handle every call and reducing benefits processing times by nearly 80%. The department is also developing predictive tools to manage call volume and create dashboards for performance monitoring, all while improving cybersecurity. Kansas Labor Secretary Amber Shultz emphasizes the focus on enhancing both claimant and staff experiences, aiming for a more agile, data-driven, and technology-embracing government system without replacing employees but empowering the existing workforce.
New survey data from VentureBeat’s July VB Pulse reveals that 39.4% of enterprise buyers plan to evaluate AI accelerators from providers other than Nvidia—such as AWS Trainium, Google TPU, AMD Instinct, Intel Gaudi, or custom ASICs—within the next year, surpassing the 25.3% considering Nvidia's upcoming Blackwell GPUs by 14 points. While Nvidia remains dominant in production, organizations are diversifying their AI accelerator strategies and increasing investment in existing infrastructure rather than rushing platform changes. Enterprises are also intensifying usage of platforms like Microsoft Azure, Google’s Gemini, and OpenAI, with improvements in operational reliability and throughput. Interest in non-Nvidia chips is particularly strong among strategic buyers and smaller companies, reflecting a shift to multi-vendor approaches and higher expectations for infrastructure performance and cost efficiency. Additionally, the market is seeing growth in open-source AI production systems and specialized neocloud providers, which offer new strategic options. Overall, enterprises are becoming more capable AI infrastructure operators, prioritizing reliability and flexibility, and preparing for future platform decisions from a position of greater maturity and optionality.
Infostealer malware is being used to replay stolen Claude session cookies into paid accounts without triggering two-factor authentication or Single Sign-On (SSO) protections. These affected accounts are self-serve with direct card billing, outside the governance of corporate identity providers, making it impossible for IT admins to revoke sessions. Anthropic has responded by notifying users, signing out stolen sessions, removing saved payment methods, and refunding fraudulent charges. However, the sessions can potentially access Gmail inboxes, conversation histories, files, and connected services authorized by users — posing a significant security risk for corporate data. Attackers often obtain these cookies from infected devices, sometimes via pirated software or spoofed download sites. Experts recommend treating AI service sessions on compromised machines as breached, revoking any tenant permissions possible, and moving heavy users to managed organizational accounts with tighter controls like OAuth grant restrictions and device-bound session credentials. Endpoint security combined with identity governance and AI policy enforcement is critical to closing this security gap.
