iiinova — where your digital self keeps evolving

Anuj Gupta @mettalrose.bsky.social · 13h

Happy to share a new pre-print from my research on AI and writing: "When "Helpful" Means LLMs Do the Writing for You: Rhetorical Offloading and Mental Models of Writing in Anthropic’s RLHF Training Data"

Abstract of the paper:

Excessive rhetorical offloading, where humans outsource writing labor to GenAI agents, may lead to the atrophying of rhetorical and cognitive skills over time. This study investigates the role of LLM algorithms in this problem. I conducted a 'critical dataset analysis' of Anthropic’s 'Reinforcement Learning from Human Feedback' (RLHF) data, which was used to align GenAI agents like Claude with its designers’ values. One of these datasets called 'base-helpful' was created by instructing crowdworkers to prompt LLMs and choose responses they found 'helpful.’ By analyzing a random sample of data where crowdworkers asked for help with writing tasks (n=65), this study found that responses where LLMs carried bulk of the rhetorical load tended to be chosen as 'helpful,' while responses routing rhetorical labor back to users tended to be rejected. This suggests that 'helpfulness' may have been operationalized in this dataset to favor excessive rhetorical offloading. It provides insights into the misalignment between the mental models of writing (MMOW) embedded in GenAI agents and those valued by writing studies and technical communication scholars. These findings highlight the need for more negotiation between contrasting MMOWs in writing technologies. To do this, I offer 'rhetorical uploading' as an alternative design goal for LLMs.

Keywords: GenAI, AI literacy, rhetorical load sharing, rhetorical offloading, cognitive offloading, rhetorical uploading, cognitive uploading, AI alignment, mental models of writing (MMOW), critical algorithm studies, critical dataset analysis

130

Anuj Gupta @mettalrose.bsky.social · 30/09/2026

Tried an experiment. The 1st is an essay by Claude, lightly edited by me. Pangram flags it 100% AI, ignoring my edits. I then transformed it with 'WriteHuman' - an AI text humanizer. Pangram passes it as 100% human. Welcome to the cat-&-mouse world of AI detectors & AI humanizers.

0114

Anuj Gupta @mettalrose.bsky.social · 28/09/2026

Great example of a workplace AI writing policy to share with your students if you teach #AILiteracy. This is an AI writing policy used at the software company called Clay which was written by their engineer Sophie Albert and shared by their co-found Varun Anand. Source: x.com/vxanand/stat...

031

Anuj Gupta @mettalrose.bsky.social · 25/09/2026

I replicated part of @GolinoHudson experiment on Pangram. Shakespeare's Sonnet 30 gets flagged as 100% AI generated, if you remove the last line. If you keep the last line, it becomes 100% human-written. Anyone using AI detection for real-world decisions must be very cautious.

054

Anuj Gupta @mettalrose.bsky.social · 25/09/2026

The MLA Framework for AI-Resilient Assessment and Grading is now public! It offers guidance for human-centered assessment cultures that support motivation, validity, agency, & connection. Grateful to have helped create it with the @mlanews.bsky.social AI & Pedagogy Working Group:

Screenshot of the first paragraph of the framework

1136

Anuj Gupta @mettalrose.bsky.social · 17/09/2026

Thrilled to meet students at Madras Christian College in India to talk about our new book Digital Literacies for Human Connection (NCTE Press), with my co-editors David Coad, Amber Buck, & Rich Shivener. Gratitude to Sam Moses, English HoD for hosting! Link to our book: my.ncte.org/s/product-de...

Poster for the event with the speakers photos and an image of their book

000

Anuj Gupta @mettalrose.bsky.social · 16/09/2026

Do you know any computer science instructors? I'm curious how they are tackling GenAI in the context of introductory Python classes. Are students being told to first learn to write basic Python code w/ GenAI, or is GenAI being integrated from the get-go in CS 101 kind of classes?

230

Anuj Gupta @mettalrose.bsky.social · 11/09/2026

What can critical AI literacy do to tackle the recent wave of fear that the news cycle on AI has created? I go back to the theory of metaphors and AI that Yasser Atef, @mahabali.bsky.social, @annamillsoer.bsky.social & I created in 2024 and argue that we resist allowing metaphors to outrun evidence.

112

Anuj Gupta @mettalrose.bsky.social · 10/09/2026

Just saw that the WAC Clearinghouse supports automated translation into many languages around the world! What a cool new design initiative This is what the page for our book Writing and Rhetoric Studies in the Loop: A GenAI Prompt Library looks like in English, Hindi, and Spanish 😃

Hindi translation of https://wacclearinghouse.org/books/wac-repository/wrstl/

Original English of https://wacclearinghouse.org/books/wac-repository/wrstl/

Spanish translation of https://wacclearinghouse.org/books/wac-repository/wrstl/

100

Anuj Gupta @mettalrose.bsky.social · 09/09/2026

Buckmaster's public statement about the ongoing Navier-Stokes equations controversy with OpenAI: cims.nyu.edu/~tristanb/st...

043

Anuj Gupta @mettalrose.bsky.social · 09/09/2026

Buckmaster's public statement about the ongoing Navier-Stokes equations controversy with OpenAI: cims.nyu.edu/~tristanb/st...

cims.nyu.edu

000

Anuj Gupta @mettalrose.bsky.social · 06/09/2026

It has been a real honor working on the @MLAnews "Task Force on AI in Research and Teaching‘s working group on AI and pedagogy" to create the "Designing AI-Resilient Assessment and Grading Framework." Please join us for our launch: webinars.mla.org/webinar/desi...

webinars.mla.org

Designing AI-Resilient Assessment and Grading - MLA Webinars

Join our free webinar to get strategies for assessing student writing and research in the AI era.

010

Anuj Gupta @mettalrose.bsky.social · 25/08/2026

Brian Gogan and I had a great time speaking with Bonni Stachowiak on the "Teaching in Higher Ed Podcast" about our new book where we theorize GenAI prompts as an emerging rhetorical genre and reflect on the important need to keep writers in the loop: teachinginhighered.com/podcast/writ...

teachinginhighered.com

GenAI Prompts as a Writing Genre, and Staying in the Loop

Anuj Gupta and Brian Gogan explore the GenAI prompt as a rhetorical genre and keeping writers in the loop on episode 636 of the Teaching in Higher Ed podcast.

110

Anuj Gupta @mettalrose.bsky.social · 23/08/2026

Ginsberg's Botanical Intelligence (BI) Disclosure Statement -August 23, 2026 If Allen Ginsberg were alive today, he may have needed to write something like the following statement while publishing "Howl:"

The image contains the following text:

Botanical Intelligence (BI) Disclosure Statement

During the preparation of this work, the author used several plants and their botanical intelligence for assistance with generating visionary content. After using these plants, the author reviewed and edited the content as needed and takes full responsibility for the content of the publication.

100

Anuj Gupta @mettalrose.bsky.social · 20/08/2026

So excited to be giving the keynote today at the 3I Conference (Inclusive and Innovative Instruction) at St. Mary's College of Maryland. I'll be talking about what faculty can do to find signals in the noise of GenAI discourse all around us. Event: smcm.edu/citl

Image contains details of the first slide of this presentation. An embedded image is the following: The image features a silver meat grinder. Going into the grinder at the top are various culturally symbolic, historical, and fun icons – such as emojis, old statutes, a computer, newspapers, an aeroplane. At the other end of the meat grinder, coming out is a sea of blue and grey icons representing chat bot responses like 'Let me know if this aligns with your vision' in a grey chat bot message symbol.

010

Anuj Gupta @mettalrose.bsky.social · 19/08/2026

Excited to talk to graduate students and faculty at the University of South Florida College of Arts and Sciences's @casatusf.bsky.social Department of Communication today about the promises and perils of GenAI in communications research.

Slide contains topic and presenter's information along with the following image: A collage featuring a vintage illustration of a woman’s head mapped with labeled sections resembling a phrenology chart. The mapped sections are overlaid by a neutral network diagram– depicting crisscrossing black lines. Two anonymous hands extend from the left side, pulling on two wires from the diagram. In the background is a panel of the Turing Machine with numerous knobs and switches, highlighting a connection between the history of computing, psychology, biology, and artificial intelligence.

010

Anuj Gupta @mettalrose.bsky.social · 14/08/2026

Looking forward to reading Timnit Gebru's new book!

120

Anuj Gupta @mettalrose.bsky.social · 21/07/2026

Thanks so much for your generous review and amplification, @annamillsoer.bsky.social :D

020

Anuj Gupta @mettalrose.bsky.social · 20/07/2026

Excited to share our new book, Writing and Rhetoric Studies in the Loop: A GenAI Prompt Library (ed. w/ Brian Gogan), out now from @WACClearingHous! We argue: GenAI prompts are a rhetorical genre that can help writers probe and align GenAI: wacclearinghouse.org/books/wac-re... #WRSTLPromptLibrary

wacclearinghouse.org

Writing and Rhetoric Studies in the Loop: A GenAI Prompt Library - WAC

Edited by Anuj Gupta and Brian Gogan Copy edited by Anuj Gupta and Brian Gogan. Designed by Mike Palmquist. Writing and Rhetoric Studies in the Loop: A GenAI Prompt Library is an open-access, peer-rev...

020

Anuj Gupta @mettalrose.bsky.social · 15/07/2026

Just got the Cover + TOC for our book Writing and Rhetoric Studies in the Loop: A GenAI Prompt Library (WAC Clearinghouse)! Brian Gogan and I frame GenAI prompts as an emerging rhetorical genre, with brilliant contributors designing prompts across 3 AI tools. More soon!

010

Anuj Gupta @mettalrose.bsky.social · 17/06/2026

I've always been curious about thought experiments as a research method used in philosophy. Anyone have any thoughts or readings to share on its role, its affordances, its constraints for knowledge making, especially across disciplines?

120

Anuj Gupta @mettalrose.bsky.social · 15/06/2026

My copy of #DigitalLiteraciesForHumanConnection is here! 🎉 My chapter “Being Human in an AI World” kicks off the GenAI section with brilliant educators answering what it means to be a human writer today. Grateful to co-editors @DCoad, Buck & Shivener, and the amazing @NCTE team!

Book cover of the book mentioned in the post

Anuj Gupta's chapter mentioned in the post

Chapter by all the editors mentioned in the post

010

Anuj Gupta @mettalrose.bsky.social · 08/06/2026

I am beyond thrilled to share that I received not one but two awards at the Computers and Writing 2026 conference!! 🏆 The Computers and Composition Hugh Burns' Outstanding Dissertation Award 🏆 The Michelle Kendrick Outstanding Digital Scholarship Award (co-authored)

Anuj Gupta standing with two awards in front of a plant.

250

Anuj Gupta @mettalrose.bsky.social · 06/06/2026

The 7Cs AI Subcommittee has eye-opening findings on what Computers & Writing scholars feel, think, and need about GenAI! Join us TODAY (June 6) 3:15–4:30pm in CHHS 294 at #CW2026 Computers and Writing 26. Led by Jason Tham & Joseph Robertshaw. Learn more: sites.google.com/view/7c-ai-t...

First slide from the group's power point presentation. Has the title and names of presenters mentioned in the post.

030

Anuj Gupta @mettalrose.bsky.social · 29/05/2026

Digital Literacies for Human Connection (NCTE Press) has an official released date! Coming to bookstores near you on July 21, 2026! Link to pre-order: my.ncte.org/s/product-de...

Poster for Digital Literacies for Human Connection (NCTE Press)

000

Anuj Gupta @mettalrose.bsky.social · 20/05/2026

"Well, as a scientist, I'm sure you know that Neils Bohr once said that everything we call real is made of thinge that cannot be regarded as real." - Amitav Ghosh, Ghost-Eye

Book cover of the bookGhost Eye by Amitav Ghosh

010

Anuj Gupta @mettalrose.bsky.social · 19/05/2026

Trying to figure out how to talk to an octopus these days. Sy Montgomery provides an excellent guide.

Book cover of Sy Montgomery's "Soul of an Octopus"

000

Anuj Gupta @mettalrose.bsky.social · 12/05/2026

TOC of Digital Literacies for Human Connection: 25 Ways to Engage Students in People-Centered Digital Practices (NCTE Press) is live! Theoretically sound, practical assignments for digital literacy with AI, social media, AR/VR & more. Pre-order now! Out in June 2026: my.ncte.org/s/product-de...

010

Anuj Gupta @mettalrose.bsky.social · 05/05/2026

One of the most difficult, not enough discussed aspects of research—something that feels simultaneously magical and systematic—is how we draw a boundary around what eventually become the objects of our study. #ResearchDiaries

010

Reposted by Anuj Gupta

David Coad, PhD @dcoad.bsky.social · 02/03/2026

PREORDERS available now - Digital Literacies for Human Connection: 25 Ways to Engage Students in People-Centered Digital Practices. @mettalrose.bsky.social @ambuck.bsky.social, Rich Shivener, and authors of 25 chapters have pour our heart and soul into this book. Check it out at #4C26!

094

Anuj Gupta @mettalrose.bsky.social · 18/02/2026

I had a great time today delivering the keynote lecture at the University of Central Florida's Knights Write Conference, titled "AI Slop: What, Why, How? - Understanding, Diagnosing, and Tackling the Flood of Low-Quality AI Content" Thanks so much to the organisers! cah.ucf.edu/writingrheto...

Image of Anuj Gupta's slide for the presentation with a photo of him and the title of the talk

030

Reposted by Anuj Gupta

Anna Mills @annamillsoer.bsky.social · 11/12/2025

Such a helpful piece on what is meant by “AI agent.” Includes context on the quest for AGI and lists characteristics associated with “agents”: “complex problem solving, tool use with potential independence from a chatbot dialog interface, autonomy, specialization, optional persona”

052

Anuj Gupta @mettalrose.bsky.social · 11/12/2025

Thanks to @annamillsoer.bsky.social, I've become more attuned to the slipperiness of the term "GenAI agent" which means so many different things to so many different people. Here is my current understanding of the term, which I'm looking forward to refine and expand: substack.com/home/post/p-...

substack.com

What is a "GenAI agent"?

How is it different or similar to ChatGPT or other AI tools?

030

Reposted by Anuj Gupta

Anna Mills @annamillsoer.bsky.social · 08/12/2025

Excited to dive into @mettalrose.bsky.social's dissertation, ‘Learning To Talk to Generative AI Chatbots’: A Corpus Study of Generative AI Prompts, an Emerging Genre for AI Literacy repository.arizona.edu/handle/10150...

repository.arizona.edu

141

Reposted by Anuj Gupta

Ted Underwood @tedunderwood.com · 08/12/2025

Sloan is always perceptive, and here I think he's put his finger on something important. It's not just that language models are good at code, but that they're good at helping people translate their problems *into* code. They both create a new opportunity surface and reduce barriers to entry there.

2212

Anuj Gupta @mettalrose.bsky.social · 04/12/2025

For those of you interested in issues of copyright and privacy with regards to GenAI, here is a very important recent legal development that would be helpful to talk about with students in AI literacy units: www.reuters.com/legal/govern...

reuters.com

OpenAI loses fight to keep ChatGPT logs secret in copyright case

OpenAI must produce millions of anonymized chat logs from ChatGPT users in its high-stakes copyright dispute with the New York Times and other news outlets, a federal judge in Manhattan ruled.

011

Anuj Gupta @mettalrose.bsky.social · 25/11/2025

For academic writers, editors, and publishers looking for a systematic overview of emerging GenAI policies in an academic field, please see Yin and Chapelle's (2025) excellent paper which provides this for the field of applied linguistics: www.sciencedirect.com/science/arti...

sciencedirect.com

A systematic examination of generative artificial intelligence (GenAI) use guidelines in applied linguistics journals

The unannounced appearance of GenAI in 2022 and the speed of its adoption by researchers have left many questions unanswered about its accepted ethica…

031

Anuj Gupta @mettalrose.bsky.social · 24/11/2025

Grateful for this feature by Georgia Jackson about my work on GenAI literacy at @CASatUSF. So grateful to be working with amazing colleagues at @joinusfenglish as we help shape what literacy means in the age of GenAI. www.usf.edu/arts-science...

usf.edu

USF faculty reframes what literacy means in the age of generative artificial intelligence

University of South Florida

000

Anuj Gupta @mettalrose.bsky.social · 19/11/2025

My PhD dissertation is now available for download to anyone in the world 🙂 ‘Learning To Talk to Generative AI Chatbots’: A Corpus Study of Generative AI Prompts, an Emerging Genre for AI Literacy repository.arizona.edu/handle/10150...

repository.arizona.edu

‘Learning To Talk to Generative AI Chatbots’: A Corpus Study of Generative AI Prompts, an Emerging Genre for AI Literacy

071

Anuj Gupta @mettalrose.bsky.social · 14/11/2025

As we wrap up 2025 and prep Spring 2026 syllabi, GenAI policies are on top of everyone's mind. I built a simple question-based heuristic to clarify my own thinking on GenAI for my class. If you’re drafting a GenAI policy, consider this to guide you: mettalrose.substack.com/p/heuristic-...

mettalrose.substack.com

Heuristic to Design Your GenAI Use Policy

A set of questions and resources to help you think as you design a GenAI Use Policy for your classes

010

Anuj Gupta @mettalrose.bsky.social · 07/11/2025

An important milestone of growing as a researcher involves transitioning from ignoring or rushing through literature reviews while reading research articles, to admiring and savouring them, especially when they are really well done.

020

Anuj Gupta @mettalrose.bsky.social · 07/11/2025

000

Anuj Gupta @mettalrose.bsky.social · 04/11/2025

Gearing up for teaching next semester at USF! ENC 3370: Writing Technologies (Generative AI Literacies for Professional Contexts) Artwork by Yutong Liu & Digit / betterimagesofai.org / creativecommons.org/licenses/by/... (1/4)

Poster for Dr. Anuj Gupta's class ENC 3370: Writing Technologies (Generative AI Literacies for Professional Contexts).

110

Anuj Gupta @mettalrose.bsky.social · 31/10/2025

The brilliant movie "Humans in the loop" that shows AI development from the perspective of data labellers in India--whose knowledge systems often come into conflict with the annotation protocols they work with--is now on Netflix! Excellent to show in AI literacy classes: youtu.be/t0bB5X98TrQ?...

youtu.be

Humans In The Loop | Official Trailer | Netflix India

YouTube video by Netflix India

010

Anuj Gupta @mettalrose.bsky.social · 31/10/2025

A really important but occluded research skill is finding and categorizing sources into a personal library that you can draw from when you write. Realizing more and more everyday that folder management in Zotero is something really worth investing time and energy into.

040

Anuj Gupta @mettalrose.bsky.social · 25/10/2025

"What why how when where" are the main question words in English which shape research inquiries in the west. Do you know of any other kinds of question words, probably from other languages or cultures which open up possibilities for asking different research questions than what English allows?

000

Anuj Gupta @mettalrose.bsky.social · 25/10/2025

After three years of researching GenAI prompt engineering, here is what I've learnt: (#SIGODC25) open.substack.com/pub/mettalro...

open.substack.com

GenAI alignment is a wicked problem

Prompt engineering is one of many ways we can tackle parts of it

000