Run browser agents for small admin tasks
A brief slide-deck style introduction to where we are now, based on recent presentations. Task: A brief slide-deck style introduction to where we are now, based on recent presentations When the Chancellor’s Office for the California Community Colleges sponsored a webinar on agentic AI recently, registration filled up at 1000, and many watched the livestream. I was honored to be on the panel tasked with provide some basic information about the challenge of agentic AI in education. I continue to add to what I shared that day. Here is a new Substack-formatted version. What is agentic AI? First, let me clarify that I’m not referring here to custom chatbots. A chatbot with special instructions sometimes gets called an “AI agent.” In this sense, custom chatbots inside of Copilot in Microsoft 365 or on platforms like PlayLab would be “agents.” Those have been discussed at length for some time. This slide deck is about a newer development. Anna Mills’ Substack is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. I’m talking about AI systems that do more than chat: agentic AI browsers take action. These are systems that move around digital environments and do things, much as a human would. They are designed to behave like a personal assistant or a coworker. Is it a browser or a chatbot? It’s a combination, and the chatbot can take control. An agentic AI browser is often an alternate version of Chrome, like Perplexity Comet or ChatGPT Atlas. There are also agentic AI browser extensions like Claude in Chrome. In both, there’s a chat window next to the browser window. The AI system can “see” the webpage. In both, the AI system can move the cursor, click, and fill out forms (if the user grants permission). Agentic AI browsers are newly easy to access As of September 2025, Perplexity offered its Comet browser and Pro subscription free to students. Plagiarism Today and Marc Watkins documented that the company directly advertised the browser for cheating. This browser is now free to all. Also last fall, OpenAI’s ChatGPT Atlas browser and Anthropic’s Claude in Chrome browser extension became available at the $20/month level. ChatGPT Atlas takes a quiz as a student An AI browser agent moves the cursor, clicks, and completes a quiz in a learning management system without any human intervention. It raises no academic integrity objections. Watch the video. Can educators detect AI agent activity in LMS logs? It’s unclear, but probably not, at least not reliably. These systems might perform some tasks faster than a human, but they often go fairly slowly. Cornell’s Center for Teaching Innovation says “...faculty will not find evidence of academic dishonesty in the Canvas logs. Unfortunately, this form of cheating is nearly impossible to detect through Canvas monitoring.” Canvas has said, “Quiz logs should not be used to validate academic integrity or identify occurrences of cheating.” Can LMS platforms block browser agents? Blackboard has said they can’t in a blog post, “The Latest Front in the Battle for Academic Integrity: Initial Thoughts in Response to the Rise of AI Agents.” Canvas spokesperson Melissa Lobel has said on LinkedIn that Canvas is “actively researching approaches for institutions and faculty to manage outside agent interaction with Canvas. This does include discussing these challenges with OpenAI and other major LLM providers.” Lockdown browsers as a stopgap Lockdown browsers limit which browser and/or extensions a student can use. They can probably block agentic AI. Some institutions are recommending these as a stopgap approach (see this LinkedIn discussion). Resources on this approach: Action Recommended: New AI Browsers Can Complete Canvas Quizzes. Here’s How to Secure Them - Joseph Brown, Colorado State University Academic Integrity Program Action Recommended: New AI Browsers Can Complete Canvas Quizzes. Here’s How to Secure Them - Joseph Brown, Colorado State University Academic Integrity Program Getting Started with Respondus LockDown Browser - Cornell’s Center for Teaching Innovation Getting Started with Respondus LockDown Browser - Cornell’s Center for Teaching Innovation University of Missouri suggests considering HonorLock for online tests University of Missouri suggests considering HonorLock for online tests Could software block agents from LMSes? Lonestar College professor Tim Mousel developed a system that has blocked some agentic browsers in LMSes.; tools: Perplexity Comet, Claude in Chrome, Claude Cowork. Human role: The original post does not say how a person stayed involved.; limitations: Significant privacy and security risks exist, as agentic browsers may follow malicious instructions or share visited site data with AI providers. Institutions require specific guidelines for faculty use. The tools are described as unreliable and slow, making them unsuitable for critical or time-sensitive tasks without human oversight. Detection of agent activity in LMS logs is currently difficult..
Tools
Published on Agent Work Atlas:
Overview
A brief slide-deck style introduction to where we are now, based on recent presentations
Workflow
A brief slide-deck style introduction to where we are now, based on recent presentations When the Chancellor’s Office for the California Community Colleges sponsored a webinar on agentic AI recently, registration filled up at 1000, and many watched the livestream. I was honored to be on the panel tasked with provide some basic information about the challenge of agentic AI in education. I continue to add to what I shared that day. Here is a new Substack-formatted version. What is agentic AI? First, let me clarify that I’m not referring here to custom chatbots. A chatbot with special instructions sometimes gets called an “AI agent.” In this sense, custom chatbots inside of Copilot in Microsoft 365 or on platforms like PlayLab would be “agents.” Those have been discussed at length for some time. This slide deck is about a newer development. Anna Mills’ Substack is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. I’m talking about AI systems that do more than chat: agentic AI browsers take action. These are systems that move around digital environments and do things, much as a human would. They are designed to behave like a personal assistant or a coworker. Is it a browser or a chatbot? It’s a combination, and the chatbot can take control. An agentic AI browser is often an alternate version of Chrome, like Perplexity Comet or ChatGPT Atlas. There are also agentic AI browser extensions like Claude in Chrome. In both, there’s a chat window next to the browser window. The AI system can “see” the webpage. In both, the AI system can move the cursor, click, and fill out forms (if the user grants permission). Agentic AI browsers are newly easy to access As of September 2025, Perplexity offered its Comet browser and Pro subscription free to students. Plagiarism Today and Marc Watkins documented that the company directly advertised the browser for cheating. This browser is now free to all. Also last fall, OpenAI’s ChatGPT Atlas browser and Anthropic’s Claude in Chrome browser extension became available at the $20/month level. ChatGPT Atlas takes a quiz as a student An AI browser agent moves the cursor, clicks, and completes a quiz in a learning management system without any human intervention. It raises no academic integrity objections. Watch the video. Can educators detect AI agent activity in LMS logs? It’s unclear, but probably not, at least not reliably. These systems might perform some tasks faster than a human, but they often go fairly slowly. Cornell’s Center for Teaching Innovation says “...faculty will not find evidence of academic dishonesty in the Canvas logs. Unfortunately, this form of cheating is nearly impossible to detect through Canvas monitoring.” Canvas has said, “Quiz logs should not be used to validate academic integrity or identify occurrences of cheating.” Can LMS platforms block browser agents? Blackboard has said they can’t in a blog post, “The Latest Front in the Battle for Academic Integrity: Initial Thoughts in Response to the Rise of AI Agents.” Canvas spokesperson Melissa Lobel has said on LinkedIn that Canvas is “actively researching approaches for institutions and faculty to manage outside agent interaction with Canvas. This does include discussing these challenges with OpenAI and other major LLM providers.” Lockdown browsers as a stopgap Lockdown browsers limit which browser and/or extensions a student can use. They can probably block agentic AI. Some institutions are recommending these as a stopgap approach (see this LinkedIn discussion). Resources on this approach: Action Recommended: New AI Browsers Can Complete Canvas Quizzes. Here’s How to Secure Them - Joseph Brown, Colorado State University Academic Integrity Program Action Recommended: New AI Browsers Can Complete Canvas Quizzes. Here’s How to Secure Them - Joseph Brown, Colorado State University Academic Integrity Program Getting Started with Respondus LockDown Browser - Cornell’s Center for Teaching Innovation Getting Started with Respondus LockDown Browser - Cornell’s Center for Teaching Innovation University of Missouri suggests considering HonorLock for online tests University of Missouri suggests considering HonorLock for online tests Could software block agents from LMSes? Lonestar College professor Tim Mousel developed a system that has blocked some agentic browsers in LMSes.
Reported result
The author reports that these tools are 'clunky, slow, and untrustworthy but still useful' for specific admin tasks. No quantitative metrics on time saved or error rates are provided. The author explicitly states they did not use AI to generate the text or ideas of the article itself, limiting the scope of AI utility to the described administrative experiments.
Human role
The original post does not say how a person stayed involved.
Limitations
Significant privacy and security risks exist, as agentic browsers may follow malicious instructions or share visited site data with AI providers. Institutions require specific guidelines for faculty use. The tools are described as unreliable and slow, making them unsuitable for critical or time-sensitive tasks without human oversight. Detection of agent activity in LMS logs is currently difficult.
Try it
I want to test agentic AI browsers for low-stakes administrative tasks. Please use [Perplexity Comet or Claude in Chrome] to perform the following: 1. Navigate to [URL of a simple form or scheduling tool]. 2. Fill out the form with [dummy data] or create a [meeting poll] for [date/time]. 3. Stop before submitting and show me the filled-out state. 4. If successful, submit and provide a screenshot of the confirmation. 5. If the agent fails or acts unpredictably, abort and report the error. Do not use real personal data or institutional credentials. Confirm with me before any final submission.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.