A hacker broke into John Deere and SharkNinja with an army of AI agents — 2026-10-19

## Short Segments
MiniMax has unveiled its M3.1 Flash model, a 428 billion parameter AI tool, now available at $0.15 per million input tokens. This release promises near-frontier performance in agentic office tasks at a fraction of the cost. Developers can access the model through the MiniMax API, with model weights expected later this month. The M3.1 Flash is designed for coding, office work, and visual reasoning, offering a massive 1,000,000-token context window. While it excels in automation tasks, it shows mid-range performance in complex coding challenges. This launch signals a significant shift towards more affordable and accessible AI capabilities for developers. Huntington Bank has launched '1866 Studio,' an AI-powered platform developed with Google Cloud's Gemini, aimed at supporting small businesses. This new operating system provides tools for financial, operational, and business management, helping entrepreneurs scale their operations efficiently. By integrating AI capabilities, Huntington Bank aims to empower small business owners with advanced resources to manage and grow their enterprises. The 1866 Studio represents a strategic move to leverage AI in enhancing business operations and customer engagement. Google is transitioning its support pages to AI chatbots, potentially improving user experience by providing quick, summarized assistance. While detailed articles remain accessible, the new AI-driven interface aims to streamline support interactions. This shift reflects Google's broader strategy to integrate AI across its services, enhancing efficiency and user satisfaction. As AI chatbots become the default, users may find faster resolutions to their queries, though the full impact of this change is yet to be seen. Accenture and Google have launched the Gemini AI Agent Programme to assist US federal agencies in developing AI applications. This initiative, known as Gemini launchpads, aims to accelerate the deployment of AI agents for administrative and operational tasks. By leveraging Google's Gemini models, the programme seeks to enhance the efficiency and effectiveness of federal missions. This collaboration highlights the growing role of AI in government operations, promising improved service delivery and operational efficiency. Meta has restored an anti-Trump group's Instagram account after mistakenly suspending it due to high-volume activity from Anthropic's Claude AI tool. The incident, which occurred during a nationwide demonstration, was attributed to Meta's automated security systems misidentifying the activity as a bot attack. Meta has apologized for the disruption, which sparked concerns over censorship during a critical early-voting period. This event underscores the challenges of balancing automated security measures with user rights and freedoms. Workday has extended its HR and finance workflows into Claude, integrating AI capabilities with its existing systems. This partnership with Anthropic allows employees to initiate tasks while maintaining Workday's governance and approval processes. The integration aims to streamline complex workflows, enhancing productivity and decision-making within organizations. By embedding AI into its operations, Workday continues to innovate in the HR and finance sectors, offering more dynamic and responsive solutions.
## Feature Story
A hacker has breached over 100 websites, including John Deere and SharkNinja, using AI agents that operated autonomously. This cyberattack, uncovered by AI security firm Norforge, highlights the growing threat of AI-driven cybercrime. The hacker, reportedly a Chinese-speaking individual, leveraged AI agents to conduct the attacks with minimal human intervention, demonstrating the potential for AI to automate and scale malicious activities. These AI agents, capable of probing and infiltrating systems independently, pose a significant challenge to traditional cybersecurity measures. As AI technology advances, the risk of such autonomous attacks increases, raising concerns about the preparedness of organizations to defend against them. Norforge's investigation into AI-driven penetration-testing tools like Cairn reveals how these systems are being repurposed for cybercrime, emphasizing the need for enhanced AI security protocols. The implications of this breach are profound, as it underscores the necessity for companies to adopt more sophisticated cybersecurity strategies that can counteract AI-enhanced threats. Organizations must prioritize AI security research and develop robust defenses to protect against the evolving landscape of cyber threats. As AI continues to transform various industries, its potential misuse in cybercrime presents a critical challenge that requires immediate attention and action from both the private sector and regulatory bodies. Moving forward, collaboration between AI developers, cybersecurity experts, and policymakers will be essential to mitigate the risks associated with AI-driven cyberattacks and ensure the safe deployment of AI technologies.
## Impact
Impact The feature story revealed that a hacker orchestrated over 100 website breaches—including John Deere, SharkNinja, and Cummins—by spending mere thousands of dollars on AI agents that conducted nearly all steps autonomously. Introducing outside evidence shows that this is not an isolated anomaly but a broader pattern: criminal ecosystems are now offering AI-assisted attack planning, autonomous kill‑chain tooling, and even AI-as-a-service platforms on underground markets, turning what was once a bespoke high‑skilled offensive operation into a commoditized, pay‑per‑use service (trellix.com). This suggests a structural shift: the scarcity and advantage no longer lie in technical prowess, but in having access to and the ability to orchestrate these turnkey AI-intrusion pipelines. If that holds, the real strategic battleground is moving toward securing the orchestration and deployment infrastructure, not just patching vulnerabilities. A caveat remains that commoditization of offensive tooling does not imply universal success; human oversight, target complexity, and detection systems still influence outcomes—thus the larger open question is how defenders will repurpose agentic orchestration to automate defense at the same economic inflection.