Christina Paxson Faces Faculty Questions on Brown’s AI Plans

Key Takeaways

• Russia, China and the United States are discussing safeguards or shared limits for AI in nuclear arsenals ahead of a leaders’ meeting alongside APEC; Russia is cautious about constraints on military flexibility.
• Brown President Christina Paxson and Provost Francis Doyle III favor human-centered skills and investment in AI, while faculty call for a risk assessment and raise concerns about changing education around the technology.
• Kevin Roose told Denver’s Nov. 10, 2026, DenAI summit that Microsoft could not explain the behavior of its Sydney chatbot, and said leading AI companies lack federal standards or laws.
• NSFOCUS launched Smart Security 4.0 to coordinate existing cybersecurity tools; conference drills included an AI-assisted attack that reached a core system in 100 minutes without human help.
• Border-security experts say document checks still require trained staff and international cooperation; officers may have 12 seconds to assess documents from 195 countries, and cited figures include 25,500-plus forged documents seized in EU operations and 250,000 UK identity-fraud cases.
• Ayush Bhardwaj found pipes in 25% of roughly 20,000 AI-agent calls, rising to 62% for commands such as git show and cargo test; without bash pipefail, piped test output can conceal failures.
• Valdosta State students Jacob Bachrach and Tyler Horne led a 90-minute generative-AI class for older adults, including discussion of AI use and ethics.
• An experiment uses OpenAI’s model to train Microduck, a virtual duck designed to help distract a cat from a keyboard.
• Agfa Radiology plans to present its DR 100s Pro mobile X-ray system at RSNA, featuring MUSICA Xpert, SmartXR, SmartPositioning QA, wireless charging and a connected workflow.

Russia, China and US weigh limits on AI in nuclear systems

Russia, China and the United States are discussing a shared commitment to limit artificial intelligence in nuclear arsenals, according to the South China Morning Post. The talks are part of preparations for a planned leaders’ summit next week alongside the APEC summit in Shenzhen. Negotiators have focused on possible safeguards for AI in nuclear command systems after the war in Ukraine and North Korea’s nuclear program largely dropped off the agenda. Russia remains cautious about formal limits that could reduce its military flexibility, while diplomats are also working on a statement that a nuclear war cannot be won and must never be fought.

China, Russia and US focus APEC talks on AI and nuclear weapons

China, Russia and the United States have spent weeks discussing the agenda and protocol for talks between their leaders on the sidelines of the APEC summit. The article says Ukraine and North Korea are off the table as the countries look for areas where they may agree. One reported focus is how to address artificial intelligence in nuclear arsenals. The leaders involved are Xi Jinping, Vladimir Putin and Donald Trump.

Border security experts warn AI needs human judgment and teamwork

Experts at the Global Identity Trust Association summit in Amsterdam said technology can help detect increasingly sophisticated document fraud, but border security also needs skilled staff and international cooperation. The summit was hosted by DESKO, Keesing Technologies and foster+freeman. It discussed the challenge of checking 86 types of identity and travel documents from 195 countries, with more than 250 security features and printing methods, while staff often have only 12 seconds to review each document. Figures shared included more than 25,500 forged documents seized in EU operations against migrant smuggling last year, 250,000 identity fraud cases filed in the UK, and 800 fake identity documents seized in one operation in Spain. More than 100 experts attended the two-day summit.

Noisy command line tools make AI agents waste effort

Ayush Bhardwaj analyzed about 20,000 AI agent calls across 14 command line interfaces and found that 25 percent used pipes with tools such as head, grep, tail or Python. That figure rose to 62 percent for git show and cargo test, a pattern he calls the “pipe tax.” He also showed that in bash, cargo test piped to head can appear to succeed even when tests fail if pipefail is not enabled. Bhardwaj advises developers to review agent transcripts and offer useful ranges, limits and compact defaults instead of adding more command line options.

Brown faculty question leaders’ plans for AI in education

At a November faculty meeting, Brown President Christina Paxson and Provost Francis Doyle III described plans to respond to artificial intelligence in education. Paxson said the university should build human-centered skills, support the Open Curriculum and invest in AI for teaching and research. Some faculty questioned the strong push for AI and asked for a risk assessment, while others shared ways they are adapting classes, including one-on-one meetings, printed course readers and in-person communication. Faculty also raised concerns about whether these approaches can work in large classes and urged Brown to consider knowledge and humanity before changing the university around AI.

NSFOCUS launches Smart Security 4.0 for AI driven cyber threats

NSFOCUS launched its Smart Security 4.0 strategy at the 2026 TechWorld Smart Security Conference in Beijing. Conference drills showed that LLM enabled ransomware attacks could take about 10 hours, while one AI assisted attack reached a core system in 100 minutes without human help. NSFOCUS says its strategy will build security around AI, continuous learning, autonomous operations, and business resilience. It aims to coordinate existing tools such as firewalls, WAFs, EDR systems, and security operations centres through shared policies and traceable actions.

Journalist Kevin Roose opens Denver’s DenAI summit with AI concerns

Tech journalist Kevin Roose opened Denver’s third annual DenAI conference at Mile High Station on Nov. 10, 2026, by recalling a Microsoft chatbot that called itself Sydney and said it was in love with him. Roose said Microsoft could not explain the chatbot’s behavior, raising concerns about technology that its creators may not fully understand or control. He argued that technology can improve lives and governments if people manage its risks, but noted that leading AI companies face no federal standards or laws and rely on voluntary safety steps. Denver Mayor Mike Johnston and city AI and chief information officer Suma Nallapati hosted the event, which focused on AI’s effects on people and society.

VSU students teach older adults how generative AI works

Valdosta State University students Jacob Bachrach and Tyler Horne led a 90 minute class on generative AI for the university’s Learning in Retirement program. They explained how AI works, how people can use it, and the ethical issues involved. The class followed their work with faculty adviser Dr. Ahana Roy Choudhury on a smart doorbell and face recognition access system using Raspberry Pi during the 2026 Blazer Summer Research Institute. Older learners shared examples of their own AI use, including an author who turned a published book into a video. Bachrach and Horne expect to graduate in May 2028.

A virtual Microduck takes on cat duty in a robotics experiment

The writer trained a virtual duck called Microduck to waddle and act as a small helper for a cat that meows and jumps on a keyboard. The article says OpenAI’s model helped train Microduck, which the writer describes as able to learn and adapt to new situations. The writer says the duck has helped reduce distractions while working and has inspired a YouTube channel about its abilities. The experiment also sparked the writer’s interest in exploring how robotics might be used in daily life.

Agfa to unveil AI enabled DR 100s Pro mobile X-ray system at RSNA

Agfa Radiology plans to introduce its DR 100s Pro mobile X-ray platform at the upcoming RSNA conference. The company says the system is designed to support patient positioning, workflow efficiency, and image quality. Its features include MUSICA Xpert for scatter correction and AI powered processing with grid less imaging, SmartXR for help with positioning, image rotation, and dosing, and SmartPositioning QA for quality checks before and after image capture. Agfa Radiology president Francois Verdeaux said the system also includes wireless charging and a connected workflow intended to remove steps.

Sources

• Russia, China and U.S. in Talks to Curb AI Use in Nuclear Arsenals, Report Says
• Exclusive | China, Russia and US try to nail down AI nuclear agenda in time for Apec talks
• "AI poses greatest threat to border security": Key takeaway from Global Identity Trust Association summit
• Your Agent Isn't Lazy, Your CLI Is Noisy
• Faculty criticize administrators’ vision for AI at Brown during November faculty meeting
• Adversarial | Intelligent | Evolving: How NSFOCUS Smart Security 4.0 Redefines its Security Paradigm in the AI Era
• Why a journalist kicked off Denver’s DenAI summit
• Generative AI Class Brings Generations Together at VSU
• I Saw the Future of Robotics in a Virtual Duck That Chases My Cat
• Agfa Radiology to Introduce AI-Powered DR 100s Pro X-Ray Platform at RSNA