Very interesting findings regarding AI assistance in scientific writing ✍️ Also pretty large scale data that should be useful for research in this space 👇
🚩 AI-assisted scientific writing faces increasing scrutiny
Can human or LLM editing fix it?
👇 240k edits on scientific abstracts expose LLM writing flaws, and show that human editing fails to fix them
cc @hsanchaita, @leadoeun27, @shocheen, @mbodhisattwa
🧵
No Digg Deeper questions have been answered for this story yet.
Particularly excited to share our insights from 240k edits that were captured through our previous incentivized scientific editing experiment.
It's fascinating how reputation, AI stigma, and experience mediate adoption and effective use of AI in science comms.
🚩 AI-assisted scientific writing faces increasing scrutiny
Can human or LLM editing fix it?
👇 240k edits on scientific abstracts expose LLM writing flaws, and show that human editing fails to fix them
cc @hsanchaita, @leadoeun27, @shocheen, @mbodhisattwa
🧵
📢 New paper where we evaluate the process of how researchers edit AI-generated scientific text by analyzing the entire editing trajectories instead of just the final edit. we study: what kind of edits they make and when, how does it change the quality of text, and a lot more 👇
🚩 AI-assisted scientific writing faces increasing scrutiny
Can human or LLM editing fix it?
👇 240k edits on scientific abstracts expose LLM writing flaws, and show that human editing fails to fix them
cc @hsanchaita, @leadoeun27, @shocheen, @mbodhisattwa
🧵