Microsoft Unveils AI Privacy Rules for Students, OpenAI & Anthropic Praise

Key Takeaways

• <ul><li>F5 launches agentless Workforce AI Security on September 15, 2026, with policy control over employee AI use and AI agent actions</li><li>Microsoft's new AI privacy rules for students require explicit consent and third-party audits, with OpenAI and Anthropic endorsing the agreement</li><li>Multi-modal world models still struggle with cross-modal consistency, especially in video predictions that conflict with real physics</li><li>SCALR AI offers a free agentic security platform through Amazon Marketplace for automated incident triage</li><li>Crusoe and Perplexity signed a multi-year AI infrastructure partnership expected to deliver cost savings</li><li>NVIDIA is considered a stronger AI stock pick than Broadcom due to faster data center growth and lower valuation</li><li>Financial Finesse's Aimee tool uses closed-loop coaching logic to avoid hallucinations, earning award recognition</li><li>An aviation safety framework called ATAL checks AI reliability before humans use it for flight planning decisions</li><li>AI consultant Mark Laurence proposes 20 hours of mandatory AI training for government leaders at a cost of $300 million over three years</li><li>CEOs at the Fortune Leadership Forum warned that focusing AI only on efficiency gains could hold companies back from real reinvention</li></ul>

F5 launches agentless AI security tool for workforce use

F5 announced the upcoming availability of F5 Workforce AI Security on September 15, 2026. This agentless offering gives organizations visibility and policy control over how employees use AI and how AI agents act on their behalf. It works across browsers, coding agents, and MCP clients without requiring another endpoint client. The tool inspects tool calls and enforces policy before risky actions occur. It integrates into existing SASE environments.

F5 adds workforce AI security to its platform

F5 (NASDAQ: FFIV) announced F5 Workforce AI Security on September 15, 2026. The new agentless tool gives enterprises control over employee AI use and AI agent actions. It provides capabilities to govern AI use, understand identity and intent, control agent actions before execution, and enforce policy in the interaction path. Kunal Anand, Chief Product Officer at F5, said it applies intent-based guardrails in the network path. The offering integrates seamlessly into existing SASE environments.

F5 rolls out Workforce AI Security for enterprises

F5 (NASDAQ: FFIV) announced upcoming availability of F5 Workforce AI Security, a new agentless offering within the F5 AI Security Platform. It gives organizations visibility and policy control across workforce AI activity, including employee AI use and actions taken by agents on users' behalf. The announcement was made on September 15, 2026. The tool is designed to cover AI interactions wherever they run.

Microsoft sets new AI privacy rules for students

Microsoft announced a new set of AI privacy rules for students across its products and services. The rules require Microsoft to obtain explicit consent from students before collecting and using their data. Students will also have more control over their data. The announcement comes as educators and policymakers grow concerned about AI use in education. The move is expected to influence other tech giants to adopt similar protections.

Microsoft agrees to AI privacy protections for students

Microsoft has committed to sweeping AI privacy rules for students in response to growing concerns about AI in schools. New York City and Los Angeles banned AI tools for a year at the start of the school year. The agreement is legally binding and requires third-party audits for compliance. OpenAI and Anthropic praised the agreement, with OpenAI calling it an important milestone. Experts say the standards could help shape conversations between schools and AI providers.

Study finds multi-modal world models still lack internal and physical consistency

Researchers studied world models that generate multiple types of outputs at once, such as video and text. They found that even when each output looks believable on its own, the model can still contradict itself across modalities. The team built a physics-based testing pipeline to measure these failures. Across 20 test settings, language tasks scored well, but video predictions often disagreed with real physics. The results suggest current unified models still struggle with reliable reasoning and physical accuracy.

Expert calls for mandatory AI training for government leaders and CEOs

AI consultant Mark Laurence proposed that the next government require 20 hours of mandatory AI training for all ministers and chief executives. His plan would cost about $300 million over three years. He said leadership and cultural change matter more than the technical side of AI. He stressed that leaders need hands-on experience to truly understand the technology. Without that experience, he believes organizations and society cannot fully adapt to the changes AI brings.

CEOs warn that AI efficiency gains risk holding companies back

Business leaders speaking at the Fortune Leadership Forum in Macau urged caution about using AI only to cut costs. They warned that focusing too much on efficiency could produce a better version of past methods rather than real reinvention. Leaders including AS Watson Group CEO Malina Ngai and BCG Greater China chair emphasized that judgment matters more as AI makes knowledge widely available. Uber CEO Dara Khosrowshahi linked job cuts of about 3,300 corporate roles to plans for lower ride prices. The discussion highlighted that AI success depends on strong leadership and strategic thinking.

SCALR AI launches free agentic security platform through Amazon Marketplace

Security Risk Advisors introduced SCALR AI, a free platform available through Amazon Marketplace. The tool is designed to automate security tasks using agentic AI. It offers an agentive workbench with AI-driven quality review for incident triage and enrichment. The platform aims to help security teams handle alerts more efficiently without license fees. The launch was reported on September 15, 2026.

Crusoe and Perplexity sign multi-year AI infrastructure partnership

Crusoe and Perplexity announced a multi-year partnership focused on AI infrastructure. The deal is expected to improve efficiency and deliver cost savings for both companies. It builds on a long history of collaboration between the two firms. The partnership will allow each company to use the other's strengths and expertise. Industry observers see the deal as a significant development in AI infrastructure.

NVIDIA beats Broadcom as the smarter AI stock pick

NVIDIA is a stronger AI stock than Broadcom because it has a bigger AI business and better profits. Its data center business is growing fast thanks to AI demand. NVIDIA's CUDA platform is widely used and its valuation is lower than Broadcom's. This makes NVIDIA a better choice for investors.

Financial Finesse AI tool wins praise as regulators warn of AI errors

Regulators are warning about AI giving wrong financial advice, a problem called hallucinations. FINRA now requires firms to control for these errors. Financial Finesse's tool Aimee was named a finalist for an award and its Chief AI Officer Edwin Jongsma was recognized as a leader. Aimee uses closed-loop coaching logic instead of the open web. This approach improved retirement readiness and emergency savings for employees.

New AI trust layer could make flight planning safer

A new framework called ATAL could make AI-assisted flight planning safer. It checks if AI outputs are reliable before humans use them for decisions. ATAL looks at semantic stability, operational consistency, and constraint validation. It maps these to a Decision Readiness Level for operators. The system was tested in an aviation study and could work in other safety-critical fields too.

Sources

• F5 expands AI Security Platform with F5 Workforce AI Security  
• F5 expands AI Security Platform with F5 Workforce AI Security
• F5 expands AI Security Platform with F5 Workforce AI Security
• Microsoft commits to sweeping AI privacy rules for students. Will other tech giants follow?
• Microsoft commits to sweeping AI privacy rules for students. Will other tech giants follow?
• One Model, Two Physical Stories: Auditing Misalignment in Multi-Modal World Modeling
• Government should make AI training mandatory for all ministers and CEO's, expert says
• CEOs warn against falling into AI’s efficiency trap: The biggest risk is ‘a better version of yesterday’
• SCALR AI offers security teams agentic security without license fees
• Crusoe and Perplexity announce multi-year AI infrastructure deal
• NVIDIA vs. Broadcom: Which AI Stock Should You Buy After Earnings?
• As Regulators Warn of AI "Hallucinations" in Financial Guidance, Financial Finesse's Closed-Loop AI Earns Fresh Industry Recognition
• Toward a Decision-Assurance Layer for AI-Assisted Flight Planning in Air Traffic Management