Remote

Mor Naaman

@Mor@hci.social

mastodon 4.7.3

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.

960 Followers

444 Following

27 Posts

Joined November 02, 2022

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 @Mor@hci.social

· 23mo ago

Online communities are increasingly impacted by GenAI -- and responding to it. In this new paper we looked at the response via analysis of AI-related rules, inspired in great part by papers on Reddit Rules from @cfiesler@hci.social et al and Community Archetypes from @bkeegan@hci.social et al. Travis has the details:

@travislloyd@hci.social

hci.social

Mor Naaman @Mor@hci.social

· 34mo ago

New paper preprint! How do Reddit moderators deal with AI-generated content? @travislloyd@hci.social tells you all about it here:

@travislloyd@hci.social

(also, follow him on Mastodon)

hci.social

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

Why is this important? It's now clear that more of our online content and communication will be generated by AI. In our previous work, we demonstrated the "Replicant Effect": as soon as the use of AI is suspected, evaluations of trustworthiness drop. https://dl.acm.org/doi/10.1145/3290605.3300469

dl.acm.org

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

In the current work, we show that not only people cannot distinguish between AI and human-written text, but that they have heuristics that can be exploited by AI, potentially leaving the poor authentic humans to suffer decreased trustworthiness evaluations. As we rush into deploying/using large language models and applications like #ChatGPT in our everyday communication, our paper asks to stop and consider the consequences as we move from Computer-mediated to AI-mediated communications (AI-MC).

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

Progress! #GPT4 nails (this) Winograd Schema that failed ChatGPT.

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

For example, use of rare bigrams & long words was associated with people thinking the profile was generated by AI. In reality, such profiles were more likely to be human-written. Use of informal speech and mentions of family were (wrongly) associated with human-written text. We also ran an experiment. Since there are predictable features that make people think a profile sounds more "human", AI can take advantage of these features/heuristics to create profiles that are "more human than human".

Mor Naaman @Mor@hci.social

· 32mo ago

Replying to

Here is me building on CMC theories to prove @ICA_CAT -worthy

Mor Naaman @Mor@hci.social

· 47mo ago

📣 📝 New paper out! #cscw2022 📝📣

Crisis informatics but when the crisis is long and sustained and you just cannot take it anymore...

Link to tweet where you could follow Marianne as well as here @marianne@hci.social

https://twitter.com/marianneaubin/status/1589637371104210944

twitter.com

Mor Naaman @Mor@hci.social

· 44mo ago

Misinformation on Misinformation. A solid (but debatable) argument.

[I wish I could say this paper was all wrong, in which case it would be... you know]

https://journals.sagepub.com/doi/10.1177/20563051221150412

journals.sagepub.com

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

In the first part of this work (lead author: Maurice Jakesch) we collected 1000s of human-written self-presentations in important contexts (dating, freelance, hospitality); created (#GPT) 1000s of AI-generated profiles; and asked 1000s of people to distinguish between them. They couldn't (success rate: ~50%). However, they were consistent: people had specific ideas about which profile was AI/human. 2/

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

The paper was a Cornell Tech & Stanford collaboration led by Maurice Jakesch (🚨on the job market in Europe🚨) and with Jeff Hancock. I will write about the origin of this NSF-funded research next. Also, more exciting AI-MC results to come soon! Paper link again: https://www.pnas.org/doi/10.1073/pnas.2208839120

pnas.org

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

Just got access to Bard, which managed the Winograd Schemas well, too. Though I really didn't need that much detail about Styrofoam 🤷‍♂️

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

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).

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

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.

Mor Naaman @Mor@hci.social

· 36mo ago

Haha Bing not quite what I had in mind

Mor Naaman @Mor@hci.social

· 43mo ago

Replying to

Neglected to add the open-access version. It's slightly older and not 100% the same text as the final PNAS paper but includes all the results/data. https://arxiv.org/abs/2206.07271

arxiv.org

Mor Naaman @Mor@hci.social

· 35mo ago

Replying to

@jbigham @andresmh yes I'm ready to join the OJ defense team, thanks for asking

Mor Naaman @Mor@hci.social

· 35mo ago

Is anyone here also on Threads? I am surprised our social computing community is not there yet. Or at least not visibly so. Wait [scary realization] are you guys STILL ON TWITTER?

I hope @andresmh@hci.social doesn't pull a Musk and ban me from hci.social for linking to the competition but if you are thread-ing find me there:

@mmoorr@www.threads.net

www

Mor Naaman @Mor@hci.social

· 35mo ago

Was fun to look at Cofacts, a knowledge-sharing community in Taiwan focused on fact-checking, and how their activities compared (computationally) to professional fact-checkers. Story here. Paper at the Journal of Trust and Safety.
Work led by Andy Zhao

https://news.cornell.edu/stories/2023/11/crowdsourced-fact-checking-fights-misinformation-taiwan

news.cornell.edu

Mor Naaman @Mor@hci.social

· 40mo ago

Replying to

👆Feel free to share with others doing important tech-driven public-interest work! (or just boost to your network).

Mor Naaman @Mor@hci.social

· 23mo ago

Replying to

@cfiesler Thanks for reading and providing that great foundation! (cc @travislloyd)

Mor Naaman @Mor@hci.social

· 35mo ago

Replying to

@jbigham @andresmh between here and Instagram I'm surprised you have time for the family