San Antonio Schools Integrate AI Tools Like MagicSchool, Diffit

Key Takeaways

• San Antonio districts including Northside and North East ISD have approved lists of AI tools like MagicSchool and Diffit for classroom use.
• Experts warn AI should not be used for grading or predicting student performance, and teacher training requires ongoing support.
• Sweet Home Central school district uses Claude from Anthropic to cut bus routing work from hours to minutes.
• Google AI Mode links to Google Shopping 57 percent of the time in product searches, according to a Findrix study of 55,198 answers.
• A University of Kansas checklist helps identify bias in AI tools across credibility, completeness, representation and accountability.
• Jonathan Swift's 1726 book Gulliver's Travels described a machine that resembles modern large language models.
• 64% of U.S. adults believe AI is developing too fast, according to a nationwide poll cited at SJSU's Responsible Innovation conference.
• Huawei Zambia launched a five-day AI training program for government leaders to support digital transformation.
• Workers Erin Maus and David Durman argued they can refuse to use AI at work based on religious rights under Title VII.
• Russian influence group Storm-1516 used AI chatbots including Google Gemini and Microsoft Copilot to spread false claims during a Texas election.

San Antonio teachers use AI as schools set guidelines

San Antonio teachers are using AI tools like MagicSchool and Diffit in classrooms. Districts such as Northside and North East ISD have approved lists of allowed tools. Experts say AI should not be used for grading or predicting student performance. Training teachers on AI takes time and money. Districts are still figuring out how to enforce rules and support educators.

One-time AI training is not enough for teachers

A one-time AI workshop does not fully prepare teachers to use AI in daily lessons. Teachers need ongoing support to decide when AI helps learning and when it does not. They need trusted colleagues to help work through real classroom problems. Time to practice and build judgment is important. A workshop can introduce rules but cannot replace continued learning.

AI could handle admin tasks so teachers focus on students

The Sweet Home Central school district in New York is using AI tools to reduce paperwork. It built a bus routing tool with Claude from Anthropic that cuts hours of work down to minutes. The district wants AI to help with reports, data entry and other admin duties. It is working with TechBuffalo and Radial Agency. The goal is to free staff so they can spend more time with students.

SJSU AI conference asks if innovation is moving too fast

San José State University held its third Responsible Innovation in AI Conference in November 2026. More than 160 students, faculty and industry leaders attended. Speakers came from companies like NVIDIA, SuperMicro and Google Cloud. The event focused on ethical development and human accountability in AI. A nationwide poll found 64% of U.S. adults believe AI is developing too fast.

Huawei sponsors AI training for Zambia's public sector

The Smart Zambia Institute and Huawei Zambia launched a five-day AI training program for government leaders. The program aims to help the public sector build and use AI solutions. It supports Zambia's digital transformation goals. Training focuses on developing skills needed to deploy AI in government work.

Workers have the legal right to refuse AI use on religious grounds

Two workers, Erin Maus and David Durman, argued they could refuse to use AI at work without penalty. They based their case on religious beliefs protected under Title VII. The right to refuse is a basic human right even if authorities can punish you. AI use can cause de-skilling and make people depend on machines. The case shows people can stand against being forced to use AI.

New checklist helps communicators spot bias in AI tools

A University of Kansas researcher Mugur Geana created a checklist to help people find bias in AI. Large language models learn from human work but miss many communities and stories. The checklist has four parts covering credibility, completeness, representation and accountability. It comes from a book called Sustainable AI that offers many views on responsible AI use. Geana says humans should always check AI output carefully.

Jonathan Swift predicted AI problems in Gulliver's Travels

Jonathan Swift wrote about a machine in his 1726 book Gulliver's Travels that looks like modern AI. In the story, a large engine with 40 handles can create books on many topics without human genius. The machine resembles today's large language models that combine words to make text. Swift showed how machines might replace human learning and creativity. His satire warned about relying too much on automated tools.

Google AI Mode mostly links to Google Shopping for products

A study by Findrix found Google AI Mode links to Google pages 46.5 percent of the time for product answers. It linked to Google Shopping 57 percent of the time in 55,198 answers. Fashion and beauty categories had the most Google links while software had the least. ChatGPT and Gemini behaved very differently from Google AI Mode. The study only tracked links and not clicks or sales.

Startups face ownership questions when AI builds their products

Startups that use AI tools to build products may not fully own what they create. The US Copyright Office says AI-generated work without enough human input may not get copyright protection. Business terms and open-source licenses can also affect ownership rights. Investors want to know a startup can legally own and transfer its technology. Keeping records of tools and human work helps protect a company's rights.

AI expert says human-focused AI needs to develop faster

Professor Choi spoke at a Stanford University conference about the fast pace of AI development. He noted that AI for AI tasks like solving math problems is advancing quickly, but AI for everyday human use is lagging. He said resources are too concentrated in one or two companies instead of spread across a healthy ecosystem. He also highlighted that current AI lacks continuous learning, meaning it cannot grow from new experiences after its initial training.

AI chatbots may boost influence campaigns during elections

An AI-generated video targeting Democratic Senate nominee James Talarico appeared online in Texas during early voting. The video falsely accused Talarico of misconduct and was viewed over 850,000 times before being removed. Russian influence group Storm-1516 used AI chatbots like Google Gemini and Microsoft Copilot to spread false claims. Experts warn this shows a new way foreign influence operations can gain credibility through AI tools during elections.

Sources

• San Antonio teachers are using AI. Schools are figuring out the rules.
• Why one-off AI training will not prepare teachers
• Can AI Take on the Grunt Work So Educators Can Focus on Student Learning?
• Are We Moving Too Fast? Takeaways from the SJSU Responsible Innovation in AI Conference NewsCenter | SJSU
• Huawei sponsors AI training for public sector in Zambia
• Opinion: Do you have the right to refuse to use AI in the workplace?
• Researcher proposes techniques for communicators to recognize, avoid inherent biases in AI
• Jonathan Swift foresaw the pitfalls of AI in Gulliver’s Travels
• Google AI Mode Sends Most Product Links to Google Shopping
• Your AI built the product. But can your startup really own it?
• "AI for artificial intelligence (AI) is developing faster than expected. But AI for humans is develo..
• AI Chatbots Could Lend Influence Campaigns New Credibility