Meta is rolling out AI-powered age estimation tools on Instagram worldwide, targeting teens aged 13 to 17. The technology identifies underage users even if they enter a fake adult birthday, placing them in Teen Accounts with built-in safeguards. Parents in the US will see notifications on Instagram this month, and Meta plans to extend the same tools to Facebook and Threads later this year. This expansion follows a $18 billion settlement related to harm claims involving teens.
Teen Accounts come with stricter settings, including limited direct messages, hidden harmful comments, and parental controls over time limits. Users aged 16 or younger cannot change settings without parent permission. Meta also continues a separate process to detect and remove children under 13. The AI analyzes clues such as birthday celebrations, school grades, and visual cues like height or bone structure.
Separately, Nvidia CEO Jensen Huang is facing controversy tied to an alleged Chinese spy who reportedly pitched AI chip deals to investors while claiming connections to Nvidia and Alibaba. The situation has drawn attention from ETF analysts who discussed potential ripple effects across the AI economy on CNBC's ETF Edge segment, aired November 9, 2026.
In cybersecurity, Cogent Security's VR-1.1 AI model outperformed Claude and GPT-based tools in vulnerability scanning, scoring 57.9% accuracy on 190 CVEs compared to 45.3% for both Claude Code and Codex. However, the model required more tokens and turns, making it less cost-efficient. Meanwhile, cybersecurity firms like Mandiant, CrowdStrike, and Kroll say they are overwhelmed by surging demand as AI-fueled attacks accelerate faster than defenders can respond.
On the education front, Cumberland County Schools is expanding use of MagicSchool AI for grades 6 through 12, offering over 80 tools for lesson planning, rubric creation, and writing feedback. Teachers retain full control by reviewing feedback before students see it. Connecticut is also preparing new rules, requiring employers using AI in hiring to disclose automated decision technology starting October 1, 2027, with violations treated as unfair trade practices.
Key Takeaways
- <ul><li>Meta expands AI age detection globally on Instagram, placing teens 13-17 in Teen Accounts with parental controls and built-in safeguards.</li><li>Meta's $18 billion settlement over teen harm claims precedes the global rollout of these protections.</li><li>Nvidia CEO Jensen Huang faces controversy linked to an alleged Chinese spy who pitched AI chip deals to investors.</li><li>Cogent Security's VR-1.1 AI model beats Claude and GPT-based tools on vulnerability scanning, scoring 57.9% accuracy versus 45.3%.</li><li>Cybersecurity firms report being overwhelmed by AI-fueled attacks, with some victims turned away due to capacity limits.</li><li>Cumberland County Schools deploys MagicSchool AI across grades 6-12 with teacher review controls in place.</li><li>Connecticut requires employers to disclose AI use in hiring decisions starting October 1, 2027.</li><li>Meta's AI age estimation tools analyze birthday celebrations, school grades, and visual cues to detect underage users.</li><li>Wirken launches open-source AI agent gateway supporting Signal, WhatsApp, Telegram, Slack, and other messaging platforms.</li><li>UConn students advise on new graduate-level AI course, debating impacts on learning, creativity, and critical thinking.</li></ul>
Meta expands AI age checks for Instagram teens globally
Meta announced it will expand AI age-estimation technology worldwide on Instagram. The tool identifies teens aged 13 to 17 and places them in Teen Accounts with built-in protections. Parents in the US will receive notifications on Instagram this month explaining the process. Meta also plans to roll out similar technology on Facebook and Threads. The move comes as Meta faces lawsuits over harm to teens and recently settled for up to $18 billion.
Instagram expands AI age detection globally for teen safety
Meta is expanding its Teen Account protections to millions more kids on Instagram worldwide. The AI age detection tools proactively look for signals to determine if users are under 18, even if they used a fake adult birthday. Parents can supervise Teen Accounts to set time limits and monitor usage. Settings cannot be adjusted by users aged 16 or younger without parent permission. This follows a separate process to detect and remove users under 13.
Meta expands Instagram AI age-guessing technology worldwide
Meta says it will use AI to find people it suspects to be teens on Instagram, even if they use an adult birthday. Teen Accounts provide stricter experiences with safeguards like limiting direct messages and hiding harmful comments. The AI tools analyze profiles for clues such as birthday celebrations, school grades, and visual cues like height or bone structure. Meta plans to continue rolling out the technology on Facebook and Threads later this year. Parents in the US will see an in-feed video on Instagram explaining the approach.
ETF Edge examines Nvidia turmoil and AI trade impact
An ETF Edge segment explores how Nvidia CEO Jensen Huang's controversy could affect the larger AI economy. Yorkville Ives partner Dan Ives and Roundhill Investments CEO Dave Mazza joined CNBC's Dominic Chu to discuss the developing story. The discussion focused on potential ripple effects across the tech trade. The segment aired on November 9, 2026.
ETF Edge discusses AI themes and Nvidia fallout
An ETF Edge segment addresses the turmoil surrounding Nvidia's CEO and its potential impact on the AI ecosystem. Roundhill Investments CEO Dave Mazza joined CNBC's Dominic Chu on the Halftime Report to share insights. The discussion covered how the situation could affect AI thematics. The segment aired on November 9, 2026.
Cogent Security VR-1.1 AI Model Beats Claude and GPT on Vulnerability Scanning
Cogent Security developed an AI model called VR-1.1 that scored 57.9% accuracy on 190 CVEs in LiveCVEBench. This outperformed Claude Code at 45.3% and Codex at 45.3% by over 12 percentage points. The model analyzes vulnerability reports, generates test code, applies patches and creates regression tests. However, VR-1.1 used more tokens and turns than competitors, making it less cost-efficient. Cogent Security was co-founded by former Abnormal Security employees and founders with Coinbase and Blackstone experience.
AI Cyberattacks Overwhelm Security Firms as Victims Turned Away
Cybersecurity firms report being overwhelmed by demand due to AI-fueled hacking. Executives at Mandiant, CrowdStrike and Kroll told NBC News that emergency cybersecurity service demand is surging. Hackers are using AI to automate attacks at a faster rate. Some victims are being turned away because security firms cannot keep up with the volume of work.
Cumberland County Schools Expands AI Tools for Classrooms
Cumberland County Schools is using MagicSchool AI platform for students and staff in grades 6 through 12. The platform offers over 80 AI-powered tools including lesson planning, rubric creation, IEP drafting and writing feedback. Teachers review all feedback before students see it, maintaining complete control. The district refined expectations for how teachers and students use AI with a school board policy committee meeting on November 3. MagicSchool is described as built specifically for K-12 education.
Connecticut Sets 2027 Rules for AI in Hiring Decisions
Connecticut Public Act 26-15 requires employers using AI in hiring or workforce management to disclose automated decision technology use starting October 1, 2027. Employers must notify employees or applicants in plain language before final decisions are made. The notice must include the technology purpose, trade name, data categories and employer contact information. Violations are treated as unfair trade practices enforceable by the Connecticut Attorney General. Employers are advised to inventory automated tools and review vendor agreements now.
Wirken Launches Open-Source AI Agent Gateway for Secure Messaging
Wirken is an open-source gateway that connects Signal, WhatsApp, Telegram, Matrix, Discord, Slack, Teams, Google Chat and iMessage. It runs each channel as a separate process holding only its own credentials. Credentials are stored in an encrypted vault and shell commands run in ephemeral containers with no network access. Every action is recorded in a signed hash-chained log for verification. Wirken is available for free and requires Docker for its sandbox feature.
UConn students to advise on new AI course
Eight University of Connecticut students were invited to help develop a new graduate-level artificial intelligence course. The discussion took place on November 6, 2026, in the Rome Ballroom. Students debated whether AI helps or hurts learning, creativity, and critical thinking. They also discussed AI risks in health care, national security, and the environment.
Alleged Chinese spy linked to Nvidia chief pitched AI chips deals to investors
Alice Huang, a Chinese national, claimed connections to Nvidia and Alibaba while seeking investment for a major data center project. She reportedly pitched AI chip deals to investors. Nvidia CEO Jensen Huang has not commented on the matter. Details about the project and investors remain limited.
Enabling Privacy-Preserving Natural Language Access to Industrial Smart Meter Data
Researchers introduced OpenMeter, a Multi-Agent System that lets non-experts query smart meter data using natural language. The tool translates plain language into code and readable insights while keeping data local to protect privacy. Studies showed accuracy improvements on limited hardware. Full orchestration is most useful where privacy requires local execution.
Sources
- Meta expands AI age checks for Instagram teens as lawsuits mount
- Instagram’s AI age detection is expanding globally
- Meta expands Instagram's AI age-guessing tech globally
- ETF Edge on how the Nvidia turmoil could impact the AI trade
- ETF Edge: AI thematics and Nvidia fallout
- Security AI Agent Outperforms GPT and Claude
- AI cyberattacks are overwhelming security firms. Some victims are being turned away.
- How Cumberland County Schools Is Bringing AI Into Classrooms
- Before the Algorithm Decides: Connecticut’s 2027 Rules for Artificial Intelligence in Employment
- Wirken: Open-source AI agent gateway
- UConn students to advise on new AI course
- Alleged Chinese spy linked to Nvidia chief pitched AI chips deals to investors
- Enabling Privacy-Preserving Natural Language Access to Industrial Smart Meter Data
Comments
Please log in to post a comment.