AI in Schools: Albany Uses Vetted Platforms, Utah Regulates Access

Key Takeaways

• Albany School District uses MagicSchool and Snorkl with strict privacy rules while some parents prefer traditional books over Chromebooks.
• Utah's Canyons School District uses a red-yellow-green stoplight protocol to regulate AI use for teachers and multilingual students.
• Pennsylvania House Bill 2788 passed 24-2 to create teacher training grants for AI literacy, privacy, and cybersecurity education.
• Florida Atlantic University leads a $5.7 million LEAP-AI program training 100+ educators in AI and literacy across 18 Broward schools.
• Penta Security released an AI cyber attack defense guide on October 6, 2026, based on visibility, defense in depth, and Zero Damage design.
• Amazon blocked Meta's Muse agent from shopping on its site; Meta is building an open agent-to-agent commerce standard with Walmart and Stripe.
• SOTA trading framework earned 18.3% return with a 1.60 Sharpe ratio over six months using Qwen3-8-27B for options trading.
• Amazon confirmed layoffs under 1,000 employees on October 8, 2026, shifting focus toward cloud and AI investment under CEO Andy Jassy.
• OpenAI's annualized revenue reported near $50 billion, below the expected $70 billion, contributing to NASDAQ dropping below 30,750.
• Cooley partners with Google to build a Gemini Enterprise AI agent for legal document redaction, expanding Big Law-Big Tech collaborations.

Albany schools use AI tools with strict rules and teacher focus

The Albany School District is using vetted AI tools like MagicSchool and Snorkl to help teachers and keep students engaged. The district says it has strict guidelines and firewalls to protect student safety and privacy. Some parents want slower AI adoption and prefer traditional books over Chromebooks. The New York State Department of Education says technology should support teachers, not replace human judgment.

Utah schools take different approaches to AI access for students

Utah public schools can now use Google Gemini, but each district handles AI differently. Canyons School District uses a stoplight protocol where red means no AI, yellow means limited use, and green means approved use. Teachers use AI to help multilingual students, plan lessons, and give feedback without sharing student names. The goal is to prepare students for real-world technology without letting AI replace their learning skills.

Pennsylvania AI training grants pass House Education Committee

Pennsylvania House Bill 2788 would create a grant program to train teachers on artificial intelligence. The bill cleared the House Education Committee with a 24-2 bipartisan vote. Schools could use grants for AI literacy, lesson planning, privacy, cybersecurity, and local policy development. Sponsor Rep. Chris Pielli says the goal is to make AI a helpful tool, not a replacement for teachers.

FAU leads $5.7M program to train teachers for AI and literacy

Florida Atlantic University is leading a three-year, $5.7 million initiative called LEAP-AI with the Florida Department of Education, Broward County Public Schools, and Miami Dade College. The program focuses on 18 high-need Broward schools and aims to improve literacy and math while training educators to use AI. It includes leadership pathways for assistant principals, teacher micro-credentials, and an apprenticeship program. At least 100 educators will earn AI and literacy credentials through the program.

Penta Security releases guide for defending against AI cyber attacks

Penta Security, a cybersecurity company based in Seoul, South Korea, released a response guide for AI-powered cyber attacks. The guide outlines three principles for cyber defense to help organizations handle these growing threats. The company says the increasing use of AI in attacks requires a new approach to protection. The guide is available for download on the company's website.

Penta Security Outlines Three Principles to Defend Against AI-Powered Cyber Attacks

Penta Security released a response guide for AI-powered cyber attacks affecting South Korean banks and financial institutions. The guide outlines three core principles: visibility, defense in depth, and Zero Damage design. Penta Security distributed the technical solution guide on October 6, 2026. The guide provides recommendations for controlling access by automated tools and detecting abnormal attack traffic. CCO Jihae Lee emphasized the importance of applying fundamental security controls without gaps.

Website Security Controls Are Blocking Personal AI Agents from Completing Tasks

Website security systems designed to block unwanted automation are now interfering with personal AI agents that shop, book flights, and complete transactions for users. Amazon has blocked Meta's Muse from browsing and purchasing on its site, while airlines like Delta and United restrict unauthorized automated access. eBay permits third-party shopping agents under certain conditions, and Yelp requires agents to access data through a licensing program. Meta is working with Walmart and Stripe to develop an open standard for agent-to-agent communication in online commerce.

SOTA Framework Achieves 18.3% Return in Stock Options Trading with AI Agents

SOTA is an agentic trading framework for structured option-strategy selection, developed by post-training Qwen3.8-27B with supervised fine-tuning and reinforcement learning. It was evaluated on options on nine large-cap U.S. equities and SPY over a six-month out-of-sample period, earning an 18.3% total return with a Sharpe ratio of 1.60 and a maximum drawdown of 8.96%. The study also found that news improves frontier-teacher trajectories but retaining news during reinforcement learning reduces out-of-sample return from 18.3% to -2.7%.

EIO-Agents Introduces Open Specification for Evaluating AI Agents

EIO-Agents is an open specification for interoperable AI agent evaluation built on two layers. The Evaluation Intelligence Ontology (EIO) provides a semantic layer through typed evidence, versioned behavioral predicates, evidence contracts, and computable derivations for PASS, REVIEW, or BLOCK decisions. The Portable Evaluation Record (PER) provides a canonical, content-addressed representation of one evaluation that preserves the evidence to decision chain. The specification aims to make AI agent evaluation more accountable and independently verifiable as agents take on greater operational responsibility.

Amazon Confirms New Round of Layoffs Affecting Under 1,000 Employees

Amazon.com Inc confirmed a new round of layoffs on October 8, 2026, affecting less than 1,000 employees across its core retail operations in the U.S. The layoffs are part of a broader restructuring strategy to realign its Stores department toward current business priorities. CEO Andy Jassy is leading efforts to streamline management levels and reduce operational costs. The move reflects a shift in focus toward cloud computing and artificial intelligence investment.

NASDAQ drops as traders sell AI stocks after Open AI revenue report

The SP500 lost ground as traders reacted to a Financial Times report on Open AI's revenue. The report showed Open AI's annualized revenue is closer to $50 billion instead of the expected $70 billion. NASDAQ fell below the support level of 30,750 to 30,800 and may drop further to 30,200 to 30,250. The Dow Jones stayed mostly flat, supported by Chevron gaining 3.5 percent.

Cooley partners with Google to build AI document redaction tool

Cooley announced a partnership with Google on Thursday to co-develop a Gemini Enterprise AI agent. The tool will help lawyers with document redaction. Cooley chose Google after identifying redaction as a strong use case for agentic artificial intelligence. The deal adds to a growing list of partnerships between Big Law and Big Tech companies.

MBody AI files patents for multi-manufacturer equipment platform

MBody AI filed two patent applications for its Orchestrator platform that manages autonomous equipment in extreme outdoor heat. The platform supports seven equipment types from three manufacturers. Its AI-as-a-Service model generates recurring monthly payments through long-term agreements. Current deployments focus on commercial cleaning, with plans to expand to delivery and security patrol tasks.

Direxion webinar explores AI investment opportunities beyond Nvidia and Anthropic

Direxion announced a webinar scheduled for October 15 about finding AI trades beyond headline names. The webinar will feature analyst Gianni Dipoce and Eliott Wellenbach discussing companies benefiting from AI spending. Companies are investing heavily in AI infrastructure, hardware, data centers and software development. The session will explore how investors can use ETFs to capitalize on the evolving AI marketplace.

Sources

• Albany schools roll out AI classroom guidance, stressing basics and teacher role
• How Utah students use AI in school depends on where they go
• Pennsylvania AI Training Grants Clear House Committee
• FAU Leads $5.7M Initiative to Boost Literacy, Prepare Educators for AI
• Penta Security Issues Response Guide for AI-Powered Cyber Attack, Outlines Three Principles for Cyber Defense
• Penta Security Issues Response Guide for AI-Powered Cyber Attack, Outlines Three Principles for Cyber Defense
• Website Security Controls Are Getting in the Way of AI Agents
• SOTA: Stock Options Trading Agents Guided by Option-Implied Return Distributions
• EIO-Agents: The Missing Semantic Layer for AI Agent Evaluation
• Amazon (AMZN) Announces New Round of Layoffs Amid AI Investment Push
• NASDAQ Index, SP500, Dow Jones Forecasts – NASDAQ Dives As Traders Sell AI Stocks
• Cooley Teams Up With Google to Co-Develop Agentic AI Redaction Tool
• MBody AI's platform brings together seven kinds of self-operating equipment from three makers.
• Beyond Just Nvidia And Anthropic For AI Trades: Opportunities Abound Beyond The Headline Names