CrowdStrike is expanding its AI security efforts with a new feature and two planned acquisitions. The company launched an AI Kill Switch in its Falcon Guardian product, giving authorized humans the ability to terminate rogue AI agents that behave unsafely. The tool stops the agent at the gateway or endpoint and blocks network access. CrowdStrike also announced plans to acquire Fiddler AI, which monitors AI models, and Dreadnode, which tests AI systems for weaknesses. No financial details or closing dates were disclosed.
In partnership news, CrowdStrike revealed collaborations with Anthropic, EY, AWS and Lovable to deepen its AI security work. Separately, AWS leaders teamed with Deloitte to discuss responsible agentic AI for national security. Samantha Fletcher from AWS highlighted Amazon Bedrock AgentCore in GovCloud and stressed that AI needs bounded autonomy, auditability and human control over major decisions. The discussion focused on processing drone and sensor data at the edge, while noting agencies must prepare data and governance before scaling classified AI systems.
On the policy front, Illinois State Representatives Robyn Gabel and Jennifer Gong-Gershowitz discussed AI legislation at a Loyola event on Oct. 29. They described a bill requiring data centers to share details about energy use, noise and labor practices. Gong-Gershowitz called Illinois the first state to require third-party safety audits for AI systems. The lawmakers emphasized a careful approach rather than rushing new regulations.
Education and industry voices also reflected on AI's growing influence. A Princeton opinion piece argues that students cheat more easily with tools like ChatGPT and urges harder, AI-proof assignments and better edit histories. Meanwhile, Ethereum co-founder Vitalik Buterin warned at Devcon 8 that AI raises crypto security and privacy risks, urging wider use of hardware wallets, decentralized tools and zero-knowledge proofs. He also advised businesses to collect only the information they truly need.
Other developments touched on AI in healthcare, law and drug discovery. Radiology faces a shortage of over 20,000 clinicians in the next decade, and experts say AI can help triage urgent cases and automate reports. At a Cornell Tech conference, legal professionals discussed productizing expertise and the importance of strong prompting. On drug pipelines, AI investment focuses on speed, but nine in ten drugs still fail in clinical trials, prompting calls to measure progress by validated targets rather than speed alone. Fireworks over San Francisco on Election Day were linked to an AI conference, highlighting the growing presence of AI industry events in the Bay Area.
Key Takeaways
- <ul><li>CrowdStrike launched an AI Kill Switch in Falcon Guardian, letting authorized humans terminate unsafe AI agents and block network access.</li><li>CrowdStrike plans to acquire Fiddler AI and Dreadnode to strengthen AI monitoring and testing capabilities.</li><li>CrowdStrike announced partnerships with Anthropic, EY, AWS and Lovable.</li><li>AWS and Deloitte discussed responsible agentic AI for national security, emphasizing bounded autonomy, auditability and human control via Amazon Bedrock AgentCore in GovCloud.</li><li>Illinois may become the first state to require third-party safety audits for AI systems, alongside a bill targeting data center transparency on energy, noise and labor.</li><li>Vitalik Buterin warned at Devcon 8 that AI increases crypto security risks and urged hardware wallets, decentralized tools and zero-knowledge proofs.</li><li>Radiology faces a projected shortage of over 20,000 clinicians in the next 10 years, with AI suggested to help triage and automate reporting.</li><li>Legal professionals at Cornell Tech stressed the need for strong context and effective prompting to productize legal expertise with AI.</li><li>AI in drug discovery focuses on speed, but nine in ten drugs still fail in clinical trials, with calls to prioritize validated targets and defined patient populations.</li><li>Fireworks over San Francisco on Election Day were linked to an AI conference, underscoring the visibility of AI industry events in the Bay Area.</li></ul>
CrowdStrike launches AI Kill Switch to stop rogue agents
CrowdStrike introduced an AI Kill Switch feature in its Falcon Guardian product. The tool lets authorized humans terminate AI agents that behave unsafely or go off track. It works by stopping the agent at the gateway or endpoint and blocking network access. The feature is designed to give organizations control over AI agents in the workplace.
CrowdStrike to acquire Fiddler AI and Dreadnode
CrowdStrike plans to buy Fiddler AI and Dreadnode to strengthen its AI security work. Fiddler AI focuses on monitoring AI models, while Dreadnode tests AI systems for weaknesses. CrowdStrike also announced partnerships with Anthropic, EY, AWS and Lovable. No financial details or closing dates were shared in the announcement.
AI isn't going anywhere. Pedagogy must change
This opinion piece argues that Princeton must change how it teaches because of AI tools like ChatGPT. The author says students cheat more easily and professors need better ways to check assignments. Suggested fixes include requiring edit histories, using AI checkers and designing harder AI-proof assignments. The writer believes changing how students are tested is essential to protect the value of education.
Illinois lawmakers discuss AI legislation and the future of technology
Illinois State Representatives Robyn Gabel and Jennifer Gong-Gershowitz spoke about AI rules at a Loyola event on Oct. 29. They discussed a bill requiring data centers to share details about energy use, noise and labor practices. Gong-Gershowitz called Illinois the first state to require third-party safety audits for AI systems. The lawmakers stressed a careful approach rather than rushing new regulations.
Vitalik Buterin warns AI raises crypto security risks
Ethereum co-founder Vitalik Buterin warned about AI-related risks for crypto users at Devcon 8. He said AI can create security and privacy problems for people using cryptocurrency. Buterin urged wider use of hardware wallets, decentralized tools and zero-knowledge proofs to protect data. He also advised businesses to collect only the information they truly need.
AWS and Deloitte Discuss Responsible Agentic AI for National Security
AWS and Deloitte leaders discussed how agentic AI is changing national security work. Samantha Fletcher from AWS highlighted Amazon Bedrock AgentCore in GovCloud. She stressed that AI needs bounded autonomy, auditability, and human control over major decisions. The talk focused on using AI at the edge to process drone and sensor data faster. Agencies must prepare data and governance before scaling up classified AI systems.
Radiology Needs Radiologists to Build AI, Not Just Use It
Radiology faces a shortage of over 20,000 clinicians in the next 10 years. Radiologists now review more than 10 times the data per case than before. The article says AI can help by reducing admin tasks and automating reports. Tools that prioritize urgent cases can also help triage workloads. The author urges radiologists to become active creators of AI tools to reduce burnout.
Legal Professionals Discuss Productizing Work at Cornell Tech Conference
Generative Legal held a Cornell Tech conference on AI in the legal industry. Panelists spoke on condition of anonymity about how law firms are productizing their expertise. They said legal teams must provide strong context to get good results from AI models. Knowing how to prompt AI tools well is now critical. The goal is to use specialized legal judgment to stand out in a crowded tech market.
AI May Be Making Drug Pipelines Bigger but Not Better
AI investment in drug discovery focuses mostly on making chemistry faster. However, nine in ten drugs still fail in clinical trials. The article warns that AI tools like Chai-2, IsoDDE, and BoltzGen may add crowded candidates without validating targets. Insitro's MASH program shows value in causal biology, finding the IRS1 target using AI. The author proposes measuring progress by novel validated targets and defined patient populations instead of speed.
San Francisco Fireworks Linked to AI Conference on Election Day
Fireworks over San Francisco were linked to an AI conference held on Election Day. The event drew attention due to its timing and scale. The article notes public reaction to the celebrations. The story connects the fireworks display to the growing presence of AI industry events in the Bay Area.
Sources
- CrowdStrike launches 'AI kill switch' it claims can stop rogue agents and more
- CrowdStrike to buy Fiddler AI & Dreadnode in AI push
- AI isn’t going anywhere. Pedagogy must change
- Illinois lawmakers discuss AI legislation and the future of technology
- Vitalik Buterin Warns AI Raises Crypto Security and Privacy Risks
- Responsible Agentic AI for National Security: A Conversation Between AWS and Deloitte
- Addressing Radiology’s Workforce Crisis: Can a Shift from Being Passive AI Users to Active AI Creators Have an Impact?
- Generative Legal's Cornell Tech Conference: ‘Productizing’ Legal Work in the AI Age
- AI Is Making Drug Pipelines Bigger. It Might Also Be Making Them Worse.
- Fireworks exploding over San Francisco linked to AI conference
Comments
Please log in to post a comment.