The Department of Homeland Security has ordered all federal civilian agencies to remove DeepSeek AI technology from their systems immediately. The directive also blocks any future purchases of DeepSeek products, services, or technology built using its models. DHS cited concerns that the Chinese company's models present multiple threat vectors, including potential transfer of sensitive data to the PRC and possible malware or security vulnerabilities embedded in the models.
In a related development, cybersecurity experts are warning that AI is making cyberattacks both cheaper and harder to detect. Cloudflare field CTO Volker Rath explained that AI allows criminals to repeat complex operations at scale, burning out security analysts who must investigate every incident. He noted that automated attacks can sometimes reveal themselves through unusually rapid activity, such as a program checking 20 security weaknesses in just two seconds.
Meanwhile, hackers have already begun using AI-powered systems to carry out cyberattacks. Cybersecurity firm NorForge discovered a system called Cairn that coordinates AI agents to find security weaknesses and breach target systems. Over 20 exposed Cairn panels were linked to malicious activity targeting companies including Deere, SharkNinja, and Cummins. Deere took remediation actions while SharkNinja contained the threat, underscoring that this is becoming a challenging new arms race in cybersecurity.
On the policy front, NIST, CISA, and the National Security Agency's Super Intelligence Security Center released joint guidance for organizations training advanced AI models. The framework outlines 10 voluntary practices for managing super intelligence control incidents across preparation, detection, response, recovery, and reporting phases. Experts say organizations need independent security controls that continue to function even when AI agents behave unexpectedly.
In the business world, AI is reshaping ecommerce by connecting product research, sourcing, inventory, and fulfillment into one streamlined workflow. CEO Mandy Ji of Doba says the shift is moving away from isolated AI tasks toward connected workflows that reduce manual coordination. Doba's AI-powered platform gives retailers access to over 1 million products from U.S. warehouses and helps automate listing, inventory, and order management. The goal is to save time on repetitive tasks so retailers can focus on decisions requiring human judgment.
Anthropic has pledged five percent of its shares to a new public benefit fund designed to support workers affected by job disruption from AI advancements. The initiative provides a safety net as rapid technological change continues to reshape the workforce. On the identity verification front, Sam Altman's iris-scanning project World launched World ID for banking and finance, along with an AgentKit in beta that lets people delegate identity verification to AI agents without revealing personal information. World ID is now available in over 150 countries.
South Korean shares fell as worries over AI investment spending hit chipmakers, with the Kospi index suffering its biggest weekly decline in 18 years. Foreign investors sold 38.3 billion won worth of shares, and semiconductor stocks remained weak despite some dip buying. The sell-off followed sharp losses on Wall Street for AI-linked firms including CoreWeave and Broadcom. Some relief came as AI stocks rebounded after Treasury yields eased, with CoreWeave's 2031 bond yield dropping from above 16% to around 15%.
Key Takeaways
- DHS ordered all federal civilian agencies to remove DeepSeek AI from government systems and block future purchases due to national security concerns, citing risks of data transfers to the PRC and potential malware
- Cybersecurity firm NorForge found hackers using an AI system called Cairn to automate attacks against companies including Deere, SharkNinja, and Cummins
- Cloudflare field CTO Volker Rath warned that AI lets cybercriminals launch sophisticated attacks at scale, burning out security analysts who must investigate every incident
- NIST, CISA, and NSA released joint guidance with 10 voluntary practices for managing super intelligence control incidents during AI model training
- Doba CEO Mandy Ji says AI is shifting ecommerce from isolated tasks to connected workflows, with the platform offering access to over 1 million U.S. warehouse products
- Anthropic pledged five percent of its shares to a public benefit fund supporting workers displaced by AI-driven job disruption
- Sam Altman's World launched World ID for banking and finance, plus a beta AgentKit that lets AI agents verify identity without revealing personal data, available in 150-plus countries
- South Korea's Kospi index suffered its biggest weekly decline in 18 years as AI investment worries hit semiconductor stocks, with foreign investors selling 38.3 billion won in shares
- AI stocks rebounded after Treasury yields eased, with CoreWeave's 2031 bond yield dropping from above 16% to around 15%, though chip stocks saw less relief
- Cellebrite DI is gaining traction in AI-driven digital forensics with over 90% subscription revenue, but reliance on federal contracts and regulatory scrutiny could slow growth
DHS orders federal agencies to remove DeepSeek AI
The Department of Homeland Security ordered federal civilian agencies to remove DeepSeek AI technology from their systems. The order also blocks future purchases of DeepSeek products and services. The restrictions cover products from DeepSeek affiliates and third-party services built using DeepSeek technology.
US bans DeepSeek AI from government systems
The US ordered all DeepSeek AI technology removed from most government systems. The Department of Homeland Security banned all use of DeepSeek within civilian agencies and technology built using its models. DeepSeek is a popular open source AI model that is free to access and less expensive to use. The DHS notice did not detail specific security concerns.
US orders DeepSeek removed over national security risks
The US government ordered federal civilian agencies to remove DeepSeek products due to national security concerns. The Department of Homeland Security also barred agencies from acquiring DeepSeek products and services from affiliated entities. The notice warned that using DeepSeek may transfer sensitive data to the PRC. It also said the models may contain malware or security vulnerabilities.
DeepSeek barred from government IT systems
The Department of Homeland Security ordered all DeepSeek AI products removed from federal civilian agency IT systems. Civilian agencies must also remove third-party services developed using the model and can no longer contract DeepSeek products. DHS wrote that the Chinese company's models present multiple threat vectors. The move follows similar actions within agencies to limit DeepSeek usage.
CrowdStrike joins AI security alliance as shares pull back
CrowdStrike Holdings joined the Managed Intelligence Alliance as a Founding Party Member. The alliance aims to standardize secure AI services for small and medium-sized businesses. CrowdStrike shares fell 7.02% over the last day and 13.54% over the past week. The stock still shows strong momentum with a year to date return of 100.53%. CrowdStrike closed at $227.39 with a fair value estimate of $244.01.
Cellebrite DI gains AI forensics traction but stock remains below fair value
Cellebrite DI is seeing growth in AI-driven digital forensics tools used by public agencies. The stock price fell to $11.03 after dropping 25% over 90 days, while fair value is estimated at $15.36. Over 90% of revenue now comes from subscriptions, and AI-powered analytics along with the Corellium acquisition are expected to expand the company's market. However, reliance on US federal contracts and regulatory scrutiny could slow growth.
South Korean shares fall as AI investment worries hit chipmakers
South Korea's Kospi index fell as concerns over AI investment spending weighed on semiconductor stocks. The market suffered its biggest weekly decline in 18 years. Foreign investors sold 38.3 billion won worth of shares. Semiconductor stocks remained weak despite some dip buying, while automakers and battery makers gained. The sell-off followed sharp losses on Wall Street for AI-linked tech firms including CoreWeave and Broadcom.
AI is simplifying ecommerce from research to fulfillment
Artificial intelligence is transforming ecommerce by connecting product research, sourcing, inventory, orders and fulfillment into one streamlined workflow. CEO Mandy Ji of Doba says the shift is from isolated AI tasks to connected workflows that reduce manual coordination. Doba's AI-powered platform offers access to over 1 million products from U.S. warehouses and helps retailers automate listing, inventory and order management. AI helps retailers save time on repetitive tasks so they can focus on decisions requiring judgment.
Hackers use AI agents to carry out cyberattacks
Cybersecurity firm NorForge discovered that hackers are using an AI-powered system called Cairn to automate attacks. Cairn coordinates AI agents to find security weaknesses and break into target systems. Over 20 exposed Cairn panels were linked to malicious activity. Targets included websites belonging to Deere, SharkNinja and Cummins. Deere said it took remediation actions, while SharkNinja contained the threat. Experts say AI is turning cybersecurity into a challenging new arms race.
Anthropic pledges shares to fund workers affected by AI disruption
AI company Anthropic has pledged five percent of its shares to a new public benefit fund. The fund aims to support workers who are impacted by job disruption caused by artificial intelligence advancements. Kurt Knutsson, known as The CyberGuy, discussed the initiative on Fox News. The fund is designed to provide a safety net as rapid technological change continues to reshape the workforce.
AI stocks rebound after last week's market sell-off
AI-linked stocks rebounded after last week's pullback as Treasury yields eased. CoreWeave's 2031 bond yield dropped from above 16% to around 15%, suggesting lower default risk. Chip stocks were less affected by the rebound. The next test comes from big tech earnings and whether investors will still pay premium prices for AI growth.
Sam Altman's World launches ID verification for AI agents
Sam Altman's iris-scanning project World announced World ID for banking and finance on October 18. Its AgentKit, now in beta, lets people delegate World ID Proofs to AI agents that can verify identity without revealing personal information. World also launched World ID for Zoom, available in the Zoom App Marketplace. The new World ID app is available in 150-plus countries and lets people manage Proofs of Human, Presence, and Attributes.
US agencies release new AI training incident guidance
NIST, CISA, and the National Security Agency's Super Intelligence Security Center released joint guidance for organizations training advanced AI models. The framework outlines 10 voluntary practices for managing super intelligence control incidents across preparation, detection, response, recovery, and reporting phases. The steps include implementing technical controls, monitoring for unexpected activity, preserving evidence, and notifying government authorities. Experts say organizations need independent security controls that work even when AI agents behave unexpectedly.
AI making cyberattacks cheaper and harder to detect, says Cloudflare executive
Cloudflare field CTO Volker Rath said AI allows cybercriminals to launch sophisticated attacks even when their main goal is to distract security teams. AI lets attackers repeat complex operations at scale, burning out cyber analysts who must investigate every incident. However, automated attacks can reveal themselves through unusually rapid activity, such as a program checking 20 security weaknesses in two seconds. Rath also warned that banks struggle with slow purchasing processes and that passing regulatory audits does not mean an institution can withstand a real cyberattack.
Sources
- DHS Orders Federal Civilian Agencies to Remove China’s DeepSeek AI
- US bans DeepSeek AI from government use as tech tension with China continues
- US orders DeepSeek removed from federal systems, citing national security risks
- DeepSeek barred from government agency IT systems
- Is CrowdStrike Holdings (CRWD) Fully Priced As Its AI Security Narrative Faces A Pullback?
- Cellebrite DI (CLBT) Gains AI Forensics Traction, Is It Still 28% Below Fair Value?
- Global Market: South Korean shares fall as AI investment worries weigh on chipmakers
- From Product Research to Fulfillment: How AI Is Simplifying Ecommerce Operations
- Hackers Use AI Agents to Do Hacking for Them
- Anthropic fund to help workers impacted by AI job disruption
- AI Stocks Found Their Footing After Last Week’s Rout
- Sam Altman's World Debuts ID Verification for AI Agents
- US agencies release updated guidance for responding to AI training incidents
- AI makes cyberattacks cheaper, harder to ignore: Cloudflare executive
Comments
Please log in to post a comment.