Shahan Ali Memon
@shahanmemon.bsky.social
2.9K followers 1.3K following 680 posts
Researching {science of AI-mediated science, metascience #SciSci, #AI4Science, generative #AI, LLMs, agents, alignment, misinformation in science} PhD @ UW. Visiting @ NYU & MSR Alum @ Carnegie Mellon Academic webpage: https://samemon.github.io
Posts Media Videos Starter Packs
Pinned
shahanmemon.bsky.social
🚨 I recently came across a weird case of #AI in #preprints, with implications for burdening #SciComm with AI-mediated #PredatoryPublishing

What did I find? Issues with the article, questionable behavior by the author, indexing problems, and AI's potential for streamlining predatory publishing.

🧵
Reposted by Shahan Ali Memon
shahanmemon.bsky.social
This.

Much of the "AI Scientist" community focuses on an instrumentalist rather taylorist approach to science. What about the goal of "knowing" and "understanding" the world?

1/15

#ScienceOfAIMediatedScience #ScAISci #AI4Science
bayesianboy.bsky.social
The vision of human-out-of-the-loop science fundamentally mistakes what knowledge production is. It aims at the production of, wait for it, knowledge: a thing that definitionally involves an epistemic agent or agents, unless you are theistic about it.
shahanmemon.bsky.social
Disclaimer: None of this exists. It's a fictional system I designed for a speculative design class two years ago of how AI might shape peer review in the near future.

With AAAI's integration of AI into peer review this year, it is starting to sound less like fiction.
shahanmemon.bsky.social
The future of peer review is here. And it already thinks your methods section could be clearer!!!
shahanmemon.bsky.social
Finally, a human editor makes the decision, because even in the age of AI, disappointment still requires that personal touch.

Our AI-mediated peer review is where efficiency meets empathy, and automation meets accountability.
shahanmemon.bsky.social
Our commitment extends beyond authors.. we take model welfare seriously. Each AI reviewer receives regular fine-tuning, positive reinforcement, and mandatory rest periods between harsh reviews.

Because burnout is not just a human problem.
shahanmemon.bsky.social
In the Revise & Resubmit stage, the AI editor evaluates whether your revisions sufficiently address the AI's earlier concerns. This recursive feedback loop optimizes accountability and fairness.
shahanmemon.bsky.social
Once the AI completes its assessment, the manuscript and its feedback are sent to the corresponding human reviewer. This ensures continuity between synthetic and organic judgment..a seamless collaboration between flesh and code. A true human-AI collaboration.
shahanmemon.bsky.social
The AI reviewer conducts the first evaluation, drawing on vast datasets of historical feedback.

Its responses mirror authentic reviewer behavior: critical, contradictory, and confidently uncertain.

It's the most authentic automated human peer review you will ever experience at a lightening speed.
shahanmemon.bsky.social
Unlike generic systems like #ChatGPT, we're proudly human- and ethics-centered. Oh yes. When you submit your paper, you pick an AI trained to mimic a specific human reviewer's style.

We give you a choice to be judged by the robot you chose yourself. We believe in your agency!
shahanmemon.bsky.social
The future of peer review is here.

Introducing Nature Instant!

We've automated rejection so you can get disappointed faster. Because why wait 6 months for reviewer 2 when you can get roasted by an #AI in seconds?

Here is how it works.

#AI4Science #PeerReview #FictionScience #FiSci #AAAI #ScAISci
Showing a fictitious system for submitting manuscripts called "Nature Instant" showing bot choices for authors to choose from as initial AI-based peer reviewers. The bots are trained on real human reviewer data.
Reposted by Shahan Ali Memon
scientificdiscovery.dev
In general I think it's hard to combat scientific misinformation when some of the best research is locked behind an academic paywall, while lots of nonsense gets published free for everyone to read in predatory journals.
Reposted by Shahan Ali Memon
benpatrickwill.bsky.social
OpenAI's VP for education recently said the company wanted to become "core infrastructure" for schools and universities. Any infrastructure, though, always depends on habituating users to its technical affordances - so I've been trying to track how it's doing that 🧵 www.nytimes.com/2025/06/07/t...
Welcome to Campus. Here’s Your ChatGPT.
www.nytimes.com
shahanmemon.bsky.social
Please share to help us increase access & support.

#diversity #phdchat #phdchat #phdvoice #academicSky
shahanmemon.bsky.social
Applying for a #PhD @ischool.uw.edu? Read 👇

Our student-run application feedback program will be open from October 20th through 1st November 2025.

Everyone applying, especially those from historically underrepresented groups or who have faced barriers in higher ed are highly encouraged to apply.
Reposted by Shahan Ali Memon
ronentk.me
1/ Exciting paper by @sh1m.bsky.social & Sophia Liu 🔥
“hypertext systems emphasize provenance, associative thinking, and user-driven meaning-making, while algorithmic systems tend to obscure process and flatten participation” arxiv.org/abs/2507.235...
shahanmemon.bsky.social
Mandatory screenshot save as the “instructor on record” 🥳
shahanmemon.bsky.social
4) On-paper quizzes to ensure we’re not losing the spirit of learning in the AI era.

Designing and #teaching a course like this is both challenging and rewarding. My goal is to teach students how to leverage #AI as a tool while not killing the spirit of problem solving and understanding.

Onwards..