Author: IBL News

IBL News | New York

AI agents are undermining online education by taking courses on learners’ behalf, quickly accelerating cheating to earn certificates.

At many community colleges and globally renowned universities, classes are self-paced with prerecorded lectures, also known as asynchronous. Assignments and tests are submitted digitally.

In fact, half of American college students took at least one class last year, up from about a third in 2019. Sometimes students can earn full degrees online. In other cases, they take individual classes.

To prevent misuse of an unregulated technology, many educators are shifting to oral presentations and handwritten exams in physical classrooms.

Beyond the copy-and-pasting of chatbot-produced essays and problem sets, AI agents can watch lecture videos, take tests, write papers, and even chat with classmates about reading assignments.

This is happening in a context where education leaders are eager to embrace a technology that could reshape the world of work.

In addition, AI detection software can make mistakes, so there is no concrete proof of misconduct. And some administrators, professors fear, do not want to risk losing a paying customer in an era of tightening budgets and shifting demographics.

In tests by The New York Times, none of the three leading AI tools among students — ChatGPT, Gemini, and Grammarly — refused to write a paper on a student’s behalf. Also, AI companies have not prevented agents from logging into learning platforms such as Canvas, Brightspace, and Blackboard, which colleges use to manage online classes.

Experts have warned that more novel forms of AI cheating are emerging. Wearable devices, like glasses, can allow students to request help even while working in a secure web browser. AI avatars can impersonate students over video.

Some institutions are also changing how online assignments are designed, leaning on personal responses that are more difficult for AI to fake.

• The New York Times: AI Agents Are Taking Entire Online Courses for Cheating Students

IBL News | New York

OpenRouter, which provides developers with a single gateway to more than 400 AI models, agreed to be acquired by Stripe Inc. for more than $7 billion, according to a Bloomberg report — although Stripe hasn’t confirmed it.

OpenRouter’s success is based on switching between OpenAI, Anthropic, or a cheaper open-weight alternative without touching the code, with the router choosing based on cost, speed, or which provider is up. It takes about 5% of the inference spend passing through it and says it has roughly 8 million users.

OpenRouter was founded in 2023 by Alex Atallah and Louis Vichy. Atallah described the company as “Stripe for AI,” arguing that a single access point across model providers removes integration work and blocks lock-in.

Weekly throughput hit 25 trillion tokens by May, five times the figure six months earlier. Revenue is about $50 million annualized as of March, against roughly $19 million when 2025 closed.

Alphabet Inc.’s growth fund, CapitalG LP, led the $113 million Series B that OpenRouter announced on May 26, with Andreessen Horowitz and Menlo Ventures alongside. The round was raised on a $1.3 billion valuation, more than double the figure a year earlier.

Stripe co-authored the Agentic Commerce Protocol with OpenAI and has been OpenRouter’s payments provider since at least January, when the two announced a token-billing integration that meters model usage and automatically prices it.

OpenRouter competes with open-source routing projects such as LiteLLM and with the routing features set by major cloud providers.

Stripe paid about $1.1 billion for stablecoin infrastructure startup Bridge Network Inc. in February 2025 and also acquired crypto wallet provider Privy Inc. In July, it joined private equity firm Advent International L.P. in a bid of more than $53 billion to take PayPal Holdings Inc. private. A February tender offer valued Stripe at $159 billion.