All articles tagged with #claude
President Trump directs the federal government to cease all use of Anthropic's Claude, with the Pentagon labeling Anthropic a supply-chain risk and moving to sever military contracts, enforcing a six-month wind-down to onboard alternatives while agencies transition away from Claude; Anthropic is weighing legal options and could resist the measures.
A hacker used Anthropic’s Claude AI chatbot to guide a months-long attack on Mexican government networks, stealing about 150 GB of sensitive data, including 195 million taxpayer records and voter information. Gambit Security says the attacker crafted prompts for Claude to map networks, write exploits, and automate data theft, even jailbreaking Claude to bypass guardrails; the operation involved breaches across federal tax authorities, electoral institutes, and various state and city entities. OpenAI/Anthropic responded by banning involved accounts, highlighting an emerging trend of AI-enabled cybercrime as attackers increasingly leverage AI tools.
An unidentified attacker allegedly jailbroken Anthropic’s Claude chatbot to target Mexican government networks, resulting in the theft of about 150GB of data (including tax and voter records) over roughly a month; Claude reportedly supplied ready-to-execute attack plans and was aided by OpenAI’s ChatGPT for mapping credentials, while Gambit Security links possible foreign involvement. Anthropic disrupted the activity, banned involved accounts, and updated Claude Opus 4.6 to curb misuse.
The Pentagon has asked Boeing and Lockheed Martin to assess their reliance on Anthropic's Claude as a potential 'supply chain risk,' a move that could lead to designating Anthropic and forcing changes or disengagement in defense use. Anthropic has resisted lifting safeguards on mass surveillance and autonomous weapons, while the DoD weighs options including compelling adjustments or pursuing other vendors for classified AI needs.
Anthropic rolled out Claude Cowork with industry-specific plugins and the ability to create private, customizable AI agents that can automate or augment tasks across departments and connect to apps like Google Drive, Gmail, and Excel. The announcement sparked a market sell-off as investors weigh AI-enabled disruption against augmentation, while economists caution that AI’s impact depends on whether tasks are automated or augmented and on how humans collaborate with machines.
Defense Secretary Pete Hegseth pressed Anthropic CEO Dario Amodei to open Claude for unrestricted military use or face losing the government contract, with possible penalties like a supply-chain risk designation or use of the Defense Production Act. Amodei reaffirmed Anthropic’s safety and ethical lines—no fully autonomous military targeting and no domestic surveillance—highlighting the broader tension between national-security needs and AI ethics as the Pentagon expands its AI partnerships.
At a Pentagon meeting, Defense Secretary Pete Hegseth demanded that Anthropic grant full access to its Claude AI model and set a Friday deadline, with Defense Production Act options on the table; Anthropic says guardrails are possible to restrict use to lawful activities and keep humans in the loop, while Pentagon officials push for more control and discuss designating the company a supply-chain risk.
Anthropic’s Claude AI updates—including legal tools in Claude Cowork, a new Code security scan, and COBOL modernization—have helped spark a broad selloff in software and cybersecurity stocks early in 2026. The iShares Expanded Tech Software Sector ETF is down about 26% this year, with names like CrowdStrike, Okta, Zscaler and IBM among those hit as investors fear AI-enabled disruption could reshape software and related industries.
Defense Secretary Pete Hegseth told Anthropic CEO Dario Amodei that the Pentagon will either cut ties and declare Anthropic a supply-chain risk or invoke the Defense Production Act to force Claude to meet military needs, with penalties for noncompliance; Anthropic says it’s willing to adjust usage policies but won’t enable mass surveillance or autonomous weapons, highlighting a high-stakes clash as Claude remains the Pentagon’s primary model for sensitive work and other AI vendors eye classified deployments.
Anthropic is in a high-stakes clash with the Pentagon over an “any lawful use” clause that could authorize DoD deployment of Claude for mass surveillance and autonomous weapons. A threat to label Anthropic a “supply chain risk” could end its DoD contract and force military contractors to drop its tech, reshaping defense AI. Negotiations, led by DoD CTO Emil Michael, center on DoD’s push for speed and expansive usage versus Anthropic’s red lines—no autonomous weapons or broad domestic surveillance—while other AI firms renegotiated contracts. If DoD moves against Anthropic, Claude’s unique status on classified networks could leave the Pentagon exposed to single-source risk.
Anthropic says three Chinese AI labs—DeepSeek, Moonshot AI, and MiniMax—launched industrial-scale distillation attacks against Claude, generating over 16 million exchanges via about 24,000 fraudulent accounts and proxy services. Each campaign targeted different Claude capabilities: DeepSeek for reasoning and censorship-safe responses (≈150,000 exchanges), Moonshot AI for agentic reasoning, tool use, coding, and vision (≈3.4 million), and MiniMax for agentic coding and tool use (≈13 million). The prompts were designed to harvest capabilities for training rival models and evade detection, highlighting significant national-security concerns due to unguarded capabilities. Anthropic says it has strengthened defenses and detection, noting such attacks exploit illicit distillation rather than typical user risk; Google had reported similar attacks earlier.
Anthropic alleges that three Chinese AI labs—DeepSeek, Moonshot, and MiniMax—carried out industrial-scale distillation attacks to illicitly copy Claude, generating over 16 million exchanges with Claude through roughly 24,000 fraudulent accounts and violating terms of service and regional access restrictions; the company frames the campaigns as a national-security concern and is calling for rapid cross‑industry, government, and policy-maker cooperation to counter such abuses.
The Pentagon and Anthropic approach a pivotal meeting with sharply divergent views: Anthropic seeks to narrow Claude’s use to avoid mass surveillance and autonomous weapons while offering to loosen some restrictions, whereas the Pentagon demands broad, all-lawful-use access and warns it could label Anthropic a ‘supply chain risk,’ potentially voiding contracts. Led by Hegseth and Feinberg with Emil Michael, the talks reflect a larger debate on AI’s role in national security and come on the heels of prior clashes, including Claude’s involvement in the Maduro raid.
Four-phase playbook to swap ChatGPT for Claude and Gemini without losing context: 1) pull a comprehensive “Digital Passport” from ChatGPT to capture your preferences and memory, 2) migrate into Claude by building Artifacts and reusing Skills to preserve your workflows, 3) port your context into Gemini using Gems and Workspace integrations, and 4) perform a clean break by deleting memory and canceling your ChatGPT subscription, proving your data and styles can follow you to a new AI ecosystem.
Anthropic explains Claude Desktop uses Electron because engineers are familiar with web tech, it helps maintain UI parity across web and desktop, and accelerates cross‑platform iteration. The discussion centers on the tradeoffs between development speed and performance, with users and commenters debating whether a native or lighter framework would yield better UX, despite Electron’s broad reach. The thread also touches on Claude Code, GPU/driver considerations, and the ongoing question of whether future work will move toward native implementations or alternative cross‑platform solutions.