<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Projects on AI Teaching Lab at Penn Carey Law</title><link>https://ai-teaching-lab.org/projects/</link><description>Recent content in Projects on AI Teaching Lab at Penn Carey Law</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sat, 09 May 2026 00:00:00 -0400</lastBuildDate><atom:link href="https://ai-teaching-lab.org/projects/index.xml" rel="self" type="application/rss+xml"/><item><title>Exam Grader</title><link>https://ai-teaching-lab.org/projects/exam-grader/</link><pubDate>Sat, 09 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/exam-grader/</guid><description>&lt;p&gt;Exam Grader is the Lab&amp;rsquo;s calibration-based grading tool for law-school essay exams. The model doesn&amp;rsquo;t grade in a vacuum — it grades against a calibration set the faculty member has already scored, learning the rubric from worked examples before scoring the rest of the stack.&lt;/p&gt;</description></item><item><title>Course Materials</title><link>https://ai-teaching-lab.org/projects/course-materials/</link><pubDate>Fri, 08 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/course-materials/</guid><description>&lt;p&gt;Course Materials is the extraction primitive that sits underneath the rest of the Teaching Tools cluster. The job is narrow and load-bearing: take the messy artifacts of a real course — PDFs, slide decks, casebook excerpts, syllabi, lecture transcripts — and turn them into clean, structured text that other tools can consume reliably.&lt;/p&gt;</description></item><item><title>Casebook Builder</title><link>https://ai-teaching-lab.org/projects/casebook-builder/</link><pubDate>Sat, 09 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/casebook-builder/</guid><description>&lt;p&gt;Casebook Builder is the Lab&amp;rsquo;s tool for curating and editing cases into custom course casebooks. It came out of a thread between Ted Ruger, Tess Wilkinson-Ryan, and Amanda Runyon about the $500 commercial casebook problem — the basic premise being that AI plus the right tooling should let faculty assemble exactly the casebook their course needs at a fraction of the cost.&lt;/p&gt;</description></item><item><title>Exam Taker</title><link>https://ai-teaching-lab.org/projects/exam-taker/</link><pubDate>Sat, 09 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/exam-taker/</guid><description>&lt;p&gt;Exam Taker is the test-subject side of the Lab&amp;rsquo;s exam work. The pipeline runs current AI models against real law-school finals — the same exams Penn Carey Law students sit for — and produces structured reports comparing model performance against the faculty&amp;rsquo;s grading rubric.&lt;/p&gt;</description></item><item><title>Course-Website Tooling</title><link>https://ai-teaching-lab.org/projects/course-website-tooling/</link><pubDate>Sat, 09 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/course-website-tooling/</guid><description>&lt;p&gt;Course-Website Tooling is the build layer underneath the Lab&amp;rsquo;s Training Materials work. It generalizes the Legal Education at Model Velocity microsite — the four-module microsite, the cloned-voice podcast pipeline, the NotebookLM audio overview, the auto-generated PDF summaries — into reusable templates that future training packages and course microsites can deploy without rebuilding from scratch.&lt;/p&gt;</description></item><item><title>Essay Creator</title><link>https://ai-teaching-lab.org/projects/essay-creator/</link><pubDate>Sat, 09 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/essay-creator/</guid><description>&lt;p&gt;Essay Creator is the Lab&amp;rsquo;s productized version of the &lt;code&gt;law-essay-generator&lt;/code&gt; skill: a faculty-facing tool for drafting law-school essay exam questions — issue spotters and fact-pattern questions — that meet assessment-science quality standards out of the box.&lt;/p&gt;</description></item><item><title>MCQ Creator</title><link>https://ai-teaching-lab.org/projects/mcq-creator/</link><pubDate>Sat, 09 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/mcq-creator/</guid><description>&lt;p&gt;MCQ Creator is the Lab&amp;rsquo;s productized version of the &lt;code&gt;law-mcq-generator&lt;/code&gt; skill: a faculty-facing tool for drafting multiple-choice exam questions that pass the structural and psychometric quality checks the assessment literature actually expects.&lt;/p&gt;</description></item><item><title>Heron — Virtual TA Slackbot</title><link>https://ai-teaching-lab.org/projects/heron/</link><pubDate>Thu, 07 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/heron/</guid><description>&lt;p&gt;Heron is a virtual TA Slackbot for law school courses. Backed by retrieval over course materials — syllabus, readings, lecture transcripts — Heron answers student questions, points to relevant materials, and escalates to faculty when something needs human judgment.&lt;/p&gt;</description></item><item><title>Class Simulations</title><link>https://ai-teaching-lab.org/projects/class-simulations/</link><pubDate>Fri, 08 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/class-simulations/</guid><description>&lt;p&gt;Class Simulations builds AI-driven exercises around the class problems students already work through each week. Instead of generic role-plays, the simulations attach to the specific hypotheticals a course assigns — so the AI counterparty pushes back the way the doctrine actually pushes back, and students get to see their reasoning tested against something that has to follow the same rules they do.&lt;/p&gt;</description></item><item><title>Newsletter Automation</title><link>https://ai-teaching-lab.org/projects/newsletter-automation/</link><pubDate>Fri, 08 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/newsletter-automation/</guid><description>&lt;p&gt;Newsletter Automation is the build side of the Lab&amp;rsquo;s monthly AI Updates newsletter. The newsletter has been publishing since February 2025; the project here is the pipeline that produces it — automated collection of candidate items through the month, an editorial workflow that surfaces what&amp;rsquo;s worth including, and Substack as the distribution layer.&lt;/p&gt;</description></item><item><title>Training Materials</title><link>https://ai-teaching-lab.org/projects/training-materials/</link><pubDate>Fri, 08 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/training-materials/</guid><description>&lt;p&gt;Training Materials produces a complete AI training package for legal-education audiences — modules, exercises, microsites, and the accompanying media — and pitches a cross-Penn version of the same package in collaboration with Bhuvnesh Jain.&lt;/p&gt;</description></item><item><title>Judiciary — Teaching the Bench</title><link>https://ai-teaching-lab.org/projects/judiciary/</link><pubDate>Thu, 07 May 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/judiciary/</guid><description>&lt;p&gt;The Lab&amp;rsquo;s judiciary workstream — the &lt;strong&gt;PennLaw-Judiciary AI Testbed&lt;/strong&gt; — provides federal judges and their chambers with structured AI access, training, and ongoing support, governed by a formal Memorandum of Understanding. The work spans use-case research, chambers-facing tools, and a continuous feedback loop with participating chambers.&lt;/p&gt;</description></item><item><title>Legal Education at Model Velocity</title><link>https://ai-teaching-lab.org/projects/legal-education-velocity/</link><pubDate>Sun, 19 Apr 2026 00:00:00 -0400</pubDate><guid>https://ai-teaching-lab.org/projects/legal-education-velocity/</guid><description>&lt;p&gt;Legal Education at Model Velocity is a shipped artifact — a four-module microsite on what AI&amp;rsquo;s pace of change means for legal education, paired with a cloned-voice podcast, a 45-minute NotebookLM audio overview, and 14 PDF summaries that distill the modules into briefing-document form.&lt;/p&gt;</description></item></channel></rss>