Using AI in Research | Stanford CS193T Guest Panel with Mark Algee-Hewitt, Megan Ma & Markus Covert
Summary
This panel explores the innovative, cross-disciplinary applications of AI across humanities, law, and bioengineering. While showcasing advanced techniques—such as using LLMs for style transfer in literary corpora, developing AI-powered deposition simulators, and building computational models of cellular life—the discussion consistently pivots to the ethical and pedagogical challenges. Key themes include the necessity of radical transparency regarding AI use, the importance of maintaining human-led critical thinking, and the need for researchers to focus on the 'design question' and architectural foresight rather than just execution efficiency.
Key takeaways
-
AI's Role in Humanities: Style and Meaning
18:22
AI can be used to analyze the relationship between style and meaning in large corpora. By performing 'style transfer' (e.g., rewriting a philosophy article as if it were a computer science article using an LLM), researchers can computationally suggest a link between a text's style and its underlying meaning, suggesting that style is not merely superficial.
-
Legal Training via Simulation
29:20
AI can revolutionize legal education by creating high-fidelity, open-source simulations. The LIFT Lab developed a deposition simulator that generates AI witnesses with adjustable personalities (e.g., combative, evasive, forgetful), allowing junior attorneys to practice high-stakes questioning and improve core legal skills.
-
The Importance of Foresight and Design
36:20
When developing AI tools, the focus should shift from merely optimizing execution (which AI handles quickly) to tackling the 'design question' and architectural foresight. The goal is to build systems that enable human thought and complex problem-solving, rather than just automating tasks.
-
Ethical Imperative: Radical Transparency
41:10
To manage the ethical quandaries of AI, the consensus among panelists is that radical transparency is necessary. This means clearly disclosing how and where AI was used, whether in academic writing, research, or professional practice, to maintain trust and facilitate necessary human conversation.
More digests on Artificial Intelligence (AI)
Technical details
-
Computational Modeling (Bioengineering)
320s
The lab built a computer model of a cell that accounts for gene content and predicts phenotype. This work falls under 'artificial life' and involves deep learning to automatically analyze microscopy image stacks and track cell individuality over time.
-
Corpus Analysis and Topic Modeling
200s
Mark Algee-Hewitt uses computational methods for 'reading at scale,' employing techniques like topic models and TSNE graphs to analyze thousands of texts and identify machine-recognized patterns across disciplines.
-
LLM Style Transfer
1102s
A method where a local LLM is prompted to rewrite content from one style/discipline (e.g., Philosophy) to mimic another (e.g., Computer Science). This tests the relationship between style and meaning.
-
Legal AI Tools
1760s
The LIFT Lab utilizes open-source tools like a deposition simulator and explores 'partner personas,' which involves the personalization of an LLM to mimic the specific behavior and intuition of a senior professional.
Mentioned resources
- Center for Spatial and Textual Analysis
- Legal Innovation Through Frontier Technology Lab (LIFT Lab)
- open-source deposition simulator
- Claude Code
Channel & topics
Watch on YouTube · Back to latest
This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.