Haha Bing not quite what I had in mind
| Homepage | https://mmoorr.github.io/www_personal/ |
| Lab | http://stech.nyc |
| Scholar | https://scholar.google.com/citations?user=IeqjwlIAAAAJ&hl=en |
| Threads | https://threads.net/mmoorr |
Mor Naaman
- 960 Followers
- 444 Following
- 176 Posts
Information Science Faculty at Cornell Tech focusing on our information ecosystem’s trustworthiness, and its impact on democracy. Taller in person. Still on Twitter at @informor.
Great opportunity at Cornell Tech for faculty on sabbatical -- or on a mission! #pitech
https://www.pi.tech.cornell.edu/pitech-faculty-impact-fellowship
I mean, how can you not love this deliriously optimistic protest (posted by @MichalFeldman9@twitter)
Just got access to Bard, which managed the Winograd Schemas well, too. Though I really didn't need that much detail about Styrofoam 🤷♂️
Our experiment asked for human/AI labels for profiles but this time included "optimized" profiles, predicted to be rated as more human. Indeed, across contexts, people rated "optimized" profiles MUCH more often as human, compared to the human or "regular" AI-generated profiles.
We used mixed methods to uncover these heuristics & computationally show that they are indeed predictive of people's evaluations... but rarely predictive of whether the text was ACTUALLY AI-generated or human-written. The figure below shows some of the features people associated with AI (network icon) or human (human), and whether this heuristic was generally wrong (red), or correct (green).
An AI test from 1971 shows you need to adjust your #ChatGPT expectations.
In 1971, Terry Winograd devised the Winograd Schemas, simple sentence understanding tasks that require the use of knowledge and commonsense reasoning. Humans understand them easily. ChatGPT... not so much.
US-based academics, looking for something meaningful AND cool* to do on your sabbatical? Here's a program for you**.
* AND in a great location
** I should add, while I don't run the program, I am pretty sure "Effective Altruism" does not count as Public Interest Tech.