ChatGPT

Introduction to ChatGPT

If you haven’t kept up with the ChatGPT developments, this NBC news article provides a useful introduction to the conversation.
Following ChatGPT’s emergence, Princeton undergraduate Edward Tian quickly developed an app detecting whether ChatGPT has been used, and several companies will monetize these AI-detecting technologies in the future. Learn more.
Read reactions from Princeton faculty in this Daily Princetonian piece.
On February 9, the McGraw Center for Teaching and Learning hosted a panel discussion on ChatGPT featuring Princeton faculty. If you have a Princeton log in, you can view the recording.
If you’re interested in the technology behind ChatGPT, check out Princeton Research Computing / PICSciE’s “How Does Chat GPT Work? An Overview of Large Language Models,” held on February 20. Part 1 of the recording is below. Parts 2 and 3 can be found here.

“Can we always tell if the writer is an algorithm?” Wai Chee Dimock asked that question about a predecessor of ChatGPT in a 2020 PMLA editor’s column. The piece was required reading for our inaugural Humanities + Data Science Institute (HDSI), which took place in January. Check it out for yourself and apply for our upcoming summer Institute (deadline: March 10).
Follow along with Amy Winecoff, data scientist at the Center for Information Technology Policy, for a workshop on Introduction to Machine Learning for the Humanities from HDSI.
If you’re a visual learner, you might like R2D3’s A Visual Introduction to Machine Learning, also assigned by the HDSI team.

CDH Faculty Director Meredith Martin recommended this essay on “dataset accountability,” forthcoming in the Ohio State Technology Law Journal.
DH Strategist Grant Wythoff drew our attention to this piece from Emily Tucker of Georgetown Law’s Center for Technology and Privacy. Tucker discusses how Center no longer uses terms like “artificial intelligence” in order “to expose and mitigate the harms of digital technologies in the lives of individuals and communities.”
If you like podcasts, "Is Ethical AI Possible?"—an episode of Sean Illing’s “The Gray Area”—features Timnit Gebru, the founder of the Distributed AI Research Institute.

Issue 3 of our research periodical Startwords takes up the dangers of the LLMs that make programs like ChatGPT possible. The issue includes pieces by three leading DH experts as well as two co-authors of a famous—and controversial—paper on LLMs: “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?”
Meredith Martin and Ryan Heuser, now research software engineer at the CDH, both contributed to last year’s Critical AI Series on Data Ontologies. You can find Meredith’s piece here, and Ryan’s here.

Last year, former CDH project manager Caterina Agostini assembled a wonderful list of resources in her report on our Machine Learning and Humanities Working Group, co-organized with the Center for Statistics and Machine Learning. Make sure to scroll down to see the full list.
The entire HDSI syllabus is available on this site. And for the especially ambitious among you, the HDSI team provided recommended readings, too!