“On the first day of A.I. Camp, a new summer program at California State University, Savannah Bosley got a hands-on introduction to Amazon Bedrock [1], a system for building artificial intelligence apps.
“I figured it wouldn’t hurt to put it on the résumé, to learn a new tool that’s maybe marketable,” said Ms. Bosley, a computer science major who graduated this year from California Polytechnic State University, a Cal State campus in San Luis Obispo.
Dozens of students attended the five-day program, which was held on the Cal Poly campus and “powered by” Amazon Web Services, the e-commerce giant’s cloud computing division. Students scooped up free swag like Amazon-branded pens and notebooks. They tackled assignments on AWS Jam, a training app where students can practice A.I. skills. They listened as Amazon employees extolled company principles like “Think Big.”
“I’ve been thinking of it like a timeshare presentation,” Ms. Bosley, 25, said. “You get the vacation — but you also have to sit through the propaganda.”
Cal State, the largest U.S. university system with 460,000 students, recently embarked on a public-private campaign — with corporate titans including Amazon, OpenAI and Nvidia — to position the school as the nation’s “first and largest A.I.-empowered” university.
One central goal is to make generative A.I. tools, which can produce humanlike texts and images, available across the school’s 22 campuses. Cal State also wants to embed chatbots in teaching and learning, and prepare students for “increasingly A.I.-driven” careers.
As part of the effort, the university is paying OpenAI $16.9 million to provide ChatGPT Edu, the company’s tool for schools, to more than half a million students and staff — which OpenAI heralded as the world’s largest rollout of ChatGPT to date. Cal State also set up an A.I. committee, whose members include representatives from a dozen large tech companies, to help identify the skills California employers need and improve students’ career opportunities.
Cal State’s growing industry ties offer a glimpse into an extraordinary shift in campus power dynamics unfolding across the United States. Some major universities are inviting tech companies, which typically supply campus computers and email, to take on a much bigger role as education thought partners, A.I. instructors and curriculum providers.
That means dominant tech companies are now helping to steer what an entire generation of students learn about A.I, and how they use it — with little rigorous evidence of educational benefits and mounting concerns that chatbots are spreading misinformation and eroding critical thinking.
Cal State’s partners, like Amazon, said they were eager to help students use a range of A.I. tools in different ways — not just use chatbots to look up answers.
“We need students to continue to build their problem-solving skills, their strategic thinking, their ability to communicate,” said Kim Majerus, Amazon Web Services’ vice president for global education.
Cal State is not alone. Last month, California Community Colleges, the nation’s largest community college system, announced a collaboration with Google to supply the company’s “cutting-edge A.I. tools” and training to 2.1 million students and faculty. In July, Microsoft pledged $4 billion for teaching A.I. skills in schools and community colleges and to adult workers.
Critics say Silicon Valley’s efforts to make A.I. chatbots integral to education amount to a mass experiment on young people.
After the introduction of ChatGPT in 2022, students began using chatbots en masse, including for cheating. Many schools are still wrestling with how and when to allow students to use the tech.
Now, as schools like Cal State work to usher in what they call an “A.I.-driven future,” some researchers warn that universities risk ceding their independence to Silicon Valley.
“Universities are not tech companies,” Olivia Guest and Iris van Rooij, two computational cognitive scientists at Radboud University in the Netherlands, recently said in comments arguing against fast A.I. adoption in academia. “Our role is to foster critical thinking,” the researchers said, “not to follow industry trends uncritically.”
Cal State’s A.I. initiative was driven in part by state officials who had heard concern from leading tech companies that local students lacked the A.I. skills the companies needed.
“They were getting complaints from California’s A.I. giants that we weren’t doing a good job in preparing our students for this evolving workforce,” said Edmund Clark, Cal State’s chief information officer.
In February, Cal State announced its sweeping A.I. initiative. From the outset, university leaders envisioned big tech companies playing a central role in the effort, enabling administrators to promote Cal State as cutting-edge, according to university planning documents obtained through records requests by a former student and shared with The New York Times.
The university would “collaborate with industry giants” to build “an A.I.-empowered higher education system that surpasses any existing model in both scale and impact,” one document said.
Some faculty members have pushed back against the A.I. effort, as the university system faces steep budget cuts. The multimillion-dollar deal with OpenAI — which the university did not open to bidding from rivals like Google — was wasteful, they added.
Faculty senates on several Cal State campuses passed resolutions this year criticizing the A.I. initiative, saying the university had failed to adequately address students using chatbots to cheat. Professors also said administrators’ plans glossed over the risks of A.I. to students’ critical thinking and ignored troubling industry labor practices and environmental costs.
Martha Kenney, a professor of women and gender studies at San Francisco State University, described the A.I. program as a Cal State marketing vehicle helping tech companies promote unproven chatbots as legitimate educational tools.
“It’s not a ‘partnership,’” Professor Kenney said. “If you switch out the product, we would never say, ‘Xerox is collaborating with San Francisco State to offer photocopiers to all the members of its community.’”
Jason Maymon, a spokesman for Cal State, said it was the university’s responsibility to prepare students and faculty for a rapidly changing world. “Like the rise of the internet, artificial intelligence is another technological revolution, and higher education can’t simply stand by and watch,” he said.
Mr. Clark, Cal State’s chief information officer, added that the university plans to teach students to critically evaluate A.I. and had hired an outside firm to assess the A.I. initiative.
He defended the OpenAI deal, saying the company offered ChatGPT Edu at an unusually low price. Still, California’s community college system landed A.I. chatbot services from Google for more than two million students and faculty — nearly four times the number of users Cal State is paying OpenAI for — for free.
The university’s new A.I. camp with Amazon was a major part of Cal State’s initiative. Administrators from many campuses submitted proposals for student projects.
Cal Poly, a school founded in 1901 as a vocational training center, held the camp. It grew out of an initiative called the Digital Transformation Hub, or DxHub, where students work with university and Amazon employees to develop apps for nonprofits and government agencies.
The aim of the camp was to offer students from every Cal State campus the chance to use A.I. to tackle campus administration challenges — such as placing transfer students in appropriate-level math courses.
Although the 80 students from 19 campuses majored in everything from computer science to zoology, they came for the same purpose: to learn A.I. career skills.
“A.I. is the next phase of life, just like the internet, which changed everything,” said Aiman Madan, a computer science major who graduated this year from Cal State San Marcos. “It’s a race, and we need to know how to get ahead.”
As the camp kicked off, employees from Cal Poly and Amazon gave students an overview of generative A.I., noting potential uses in fields like business and medicine. Ryan Matteson, DxHub’s technology director, also warned students about risks like A.I. bias and environmental costs.
“Make sure you are laser-focused on actually trying to solve a real problem for real human beings and not just chasing shiny technology,” Mr. Matteson added.
Student teams then competed on a challenge to create a video game gatekeeper character who would vet players trying to enter a mystical world.
Dianella Sy’s team instructed an A.I. system called Amazon Nova to give the team’s gatekeeper “an assertive personality in a strict and stern tone.”
“I never built an A.I. before,” said Ms. Sy, 20, a computer science major at Cal State Fullerton.
Later in the week, students watched a futuristic video from an A.I. education start-up called Matter and Space. The reel showed Native American young people wearing A.I. glasses and tracking devices that continually monitored their activities and mood.
Then Paul J. LeBlanc, a co-founder of the start-up, told students that A.I. technologies would soon diagnose illnesses faster and more accurately than doctors, displacing human expertise.
Some students said they were reminded of the dystopian show “Black Mirror.”
“I don’t want A.I. to create more inequalities and disparities,” said Charles Walker Cano, a biology major at Stanislaus State who plans to become a physician.
Finally, students dived into group projects to solve campus administrative problems. One team assigned to streamline procurement procedures used A.I. to quickly create a prototype for an app that could automatically score vendor proposals on how well they fit university criteria.
Some students said they appreciated the chance to tackle concrete problems for actual clients. Among them was Arash Peighambari, who recently graduated from Cal State San Marcos with a master’s in computer science.
During graduate school, he had worked on a research project building an A.I. system to detect problems with power grids.
“This is more client and product and commercial-oriented,” Mr. Peighambari said of the AWS camp.” [2]
1. What does Amazon Bedrock do?
Amazon Bedrock provides a way to build and scale generative AI applications by giving access to a wide selection of foundation models through a single API. It simplifies development by offering tools for data management, security, and agent creation, which allow users to build applications for tasks like data analysis, content generation, and workflow automation. Key features include security and privacy controls, agent builders for automating complex tasks, and knowledge bases for building applications with private data.
Key features and capabilities
Access to models:
Provides access to a diverse set of foundation models from various providers through a single API, making it easy to choose the right model for a specific task.
Agent builder:
Allows developers to create agents to automate complex, multi-step tasks such as data analysis and software development by breaking them down into manageable components.
Data automation:
Offers a generative AI-powered capability for building applications that automate workflows involving multimodal data like text, images, audio, and video.
Knowledge bases:
Enables developers to build applications that can retrieve and generate information from their own private data sources, improving the accuracy and relevance of responses.
Security and privacy:
Ensures that customer data used for training or inference is private and not used to improve the base models. It also offers features like data encryption and private connectivity to AWS.
Guardrails:
Provides a way to implement safeguards to filter content, prevent harmful outputs, and detect hallucinations, ensuring more reliable and trustworthy AI responses.
Simplified development:
Offers tools and a unified platform, like the SageMaker Unified Studio [3], that simplify the process of building, customizing, and deploying generative AI applications for both technical and non-technical users.
2. Big Tech Makes Cal State Its A.I. Training Ground. Singer, Natasha; Cheung, Philip. New York Times (Online) New York Times Company. Oct 26, 2025.
3. SageMaker Unified Studio is a single, integrated development environment that provides a unified platform for data and AI workflows, enabling users to access and work with data using analytics services like SQL [4] and ETL [5], and build and deploy machine learning and generative AI applications. It streamlines the end-to-end process from data preparation to model deployment by connecting and orchestrating various AWS services in one place.
Key functionalities
Data and analytics:
It integrates with services like Amazon EMR, AWS Glue, and Amazon Athena to process data. Users can write SQL queries, build visual ETL (extract, transform, load) flows, and access a data and model catalog in a unified JupyterLab experience.
Machine learning:
It provides tools for the entire machine learning lifecycle, including data preparation, training, fine-tuning, and deployment of models.
Generative AI:
It allows developers to build generative AI applications by providing access to Amazon Bedrock models and services. Users can create chat agents, knowledge bases from their own data, and implement guardrails to ensure responsible AI practices.
Workflow orchestration:
It allows for the creation and management of complex workflows, such as chaining together visual ETL jobs and query books, which can then be automated using unified scheduling.
Security and governance:
The Studio ensures security and governance through features like fine-grained access control and responsible AI guardrails.
4. SQL, or Structured Query Language, is a standard programming language for interacting with relational databases. It is used to manage and manipulate the structured data that is stored in these databases in tables of rows and columns.
Primary functions of SQL
SQL statements are used for a variety of tasks, categorized into several main types of commands:
Querying data: The SELECT command is a Data Query Language (DQL) command used to retrieve specific information from a database.
Manipulating data: Data Manipulation Language (DML) commands are used to write or modify records.
INSERT INTO: Inserts new records into a table.
UPDATE: Modifies existing records in a table.
DELETE: Removes records from a table.
Defining data: Data Definition Language (DDL) commands are used to create and modify the database structure itself.
CREATE TABLE: Creates a new table.
ALTER TABLE: Adds, deletes, or modifies columns in an existing table.
DROP TABLE: Deletes an entire table.
Controlling access: Data Control Language (DCL) commands manage database user permissions. GRANT gives a user permission to perform an action, while REVOKE removes those permissions.
How an SQL command works
When a user sends an SQL query, the database system processes it in a series of steps:
Parsing: The query is broken down into parts and checked for correct syntax. It also verifies that the user has authorization for the requested action.
Optimization: The query processor creates an efficient execution plan for retrieving or modifying the data.
Execution: The plan is executed by the storage engine, which accesses or modifies the data stored on the physical disk.
Output: The results are returned to the user or application.
5. ETL stands for Extract, Transform, and Load, a data integration process that collects data from multiple sources, cleans and converts it into a usable format, and then loads it into a central repository like a data warehouse or data lake. This process provides a unified, clean, and organized dataset for reporting, analysis, and machine learning.
The three stages of ETL
Extract:
Data is gathered from various sources, such as databases, applications, and flat files.
Transform:
The extracted data is cleaned, validated, and converted into a consistent format. This can involve removing errors, applying business rules, and reformatting the data to match the target system's schema.
Load:
The transformed data is loaded into a target data store, such as a data warehouse, data lake, or other database.
How ETL is used
Data integration:
It consolidates data from disparate systems into a single, unified view.
Business intelligence:
It prepares data for generating reports, dashboards, and other business intelligence tools.
Analytics and machine learning:
It creates a structured dataset that serves as a foundation for data analytics and machine learning initiatives.
Data quality:
It improves data accuracy and consistency by cleaning and validating it before it's stored.
Komentarų nėra:
Rašyti komentarą