Since it debuted last fall, ChatGPT—a large language model (LLM) developed by the company OpenAI—has been a major source of debate and curiosity in the academic context and beyond.
Whereas some commentators have reflected on the potential of artificial intelligence (AI) as an educational tool, others have emphasized how ChatGPT and new tools like it pose a threat to academic integrity.
Media coverage of ChatGPT presents an opportunity for all of us in the Princeton community to commit to better understanding the methods and technologies that impact our daily lives.
At the CDH, these questions drive our work every day. Last month, we convened our inaugural Humanities + Data Science Institute (HDSI) for faculty and designated graduate students, with the goal of giving humanists the knowledge and confidence to participate in and contribute to discussions about data science and machine learning. We included a few of the readings in our list of resources below. If you’d like to dive deeper with us, we are now accepting applications for our summer Institute.
CDH-affiliated researchers are actively engaged in scholarship at the intersections of machine learning and the humanities. 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?” The issue is a must-read for those concerned about the ethics of AI. Moreover, Faculty Director 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.
Below is a list of additional resources on ChatGPT; the humanities and machine learning; and more. We hope they help you learn more about these issues—whether you are totally new to AI or have been thinking about it for a long time.
And, of course, we invite you to participate in our events and initiatives and to reach out to us individually if you’d like to share ideas, questions, or concerns.
We can’t wait to collaborate with you!
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.
- Princeton University Library is hosting “ChatGPT in Higher Education: Basics” on February 16 at 1:00 pm. Register for the Zoom.
- Read reactions from Princeton faculty in this Daily Princetonian piece.
- 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” on February 20 at 4:30 pm. Some knowledge of Python is helpful but not required. Sign up now!
Introduction to Machine Learning and the Humanities
- “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 HDSI. Check it out for yourself!
- 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 January’s HDSI.
- If you’re a visual learner, you might like R2D3’s A Visual Introduction to Machine Learning, also assigned by the HDSI team.
Law, Ethics, and AI
- Meredith 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.
- 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!