Lisa L. Ouellette
LightNews LightNews
banner
patentscholar.bsky.social
Lisa L. Ouellette
@patentscholar.bsky.social
6.8K followers 170 following 66 posts
Prof at @StanfordLaw.bsky.social, Senior Fellow at @SIEPR.bsky.social, physics PhD. Researching IP & innovation. Coauthor of free patent casebook: patentcasebook.org
Posts Replies Media Videos
patentscholar.bsky.social
Also, these are not the only types of users who might want to access Patent Center! Understanding a patent's prosecution history is relevant for businesses, academics, lawyers who aren't registered with the USPTO, etc.
Self-enrollment form for Patent Center states "Please provide the following information in order to complete the self-enrollment process for secure access to the Patent Center" and gives only 2 options: "Registered patent practitioner" or "Independent Inventor"
October 12, 2025 at 4:47 PM
patentscholar.bsky.social
New article with @marklemley.bsky.social on how genAI facilitates academic plagiarism, why this shouldn't become a copyright/legal problem, and what universities should do about it (forthcoming in U Chi L Rev Online): papers.ssrn.com/abstract=539...
The figure shows a Venn diagram with three overlapping circles labeled “Copyright Infringement,” “Plagiarism,” and “Bad Scholarly Practices.” It maps common behaviors to the regions where these circles intersect: copying expression without attribution (in all three); copying ideas or facts without attribution (plagiarism + bad scholarly practice); copying expression with attribution but beyond what’s needed for a research purpose (copyright infringement + bad scholarly practice); copying ideas or facts with attribution to unreliable sources (bad scholarly practice only); and copying expression with attribution for a research purpose (copyright infringement only, but will generally be fair use).
August 20, 2025 at 11:55 PM
patentscholar.bsky.social
Just posted a new draft with Amy Motomura, Jason Reinecke & Jonathan Masur: Can AI Hold Office Hours? Feedback welcome and appreciated! papers.ssrn.com/sol3/papers....
GenAI tools were given a patent casebook and asked 185 preregistered questions (based on real student Qs & Qs in the casebook); answers were double-coded on 3-level scale. Figure shows percentage of questions in each category that received each quality score for each model. Across all categories, a substantial number of responses were unacceptable in the sense of being harmful for learning (26% for GPT, 14% for Claude, and 31% for NotebookLM), and many more responses failed to fully answer the question or had minor errors (25% for GPT, 31% for Claude, and 32% for NotebookLM).
March 5, 2025 at 7:03 PM
patentscholar.bsky.social
Anyone know why the Federal Circuit isn't an option when selecting a jurisdiction in Lexis+ AI?
Screenshot from Lexis+AI showing federal jurisdictions including all circuits except the Federal CIrcuit
January 19, 2024 at 3:51 PM