• kth.se
  • Student web
  • Intranet
  • kth.se
  • Student web
  • Intranet
Login
DD2301 HT23 (51310)
Seminar Period 2
Skip to content
Dashboard
  • Login
  • Dashboard
  • Calendar
  • Inbox
  • History
  • Help
Close
  • Min översikt
  • DD2301 HT23 (51310)
  • Assignments
  • Seminar Period 2
2023 HT
  • Home
  • Syllabus
  • Assignments
  • Course Evaluation

Seminar Period 2

  • Due No due date
  • Points 0
  • Submitting a file upload
  • File types pdf

Generative AI and the Law

Generative AI in the past year has become a "useful" technology - especially the development of LLMs, voice synthesis, image generation,.... . Now governments and the legal world are in catch-up mode to decide how existing legislation can be used and/or be adapted to cope with this new technology. Within many countries and regions, there is lively debate and discussion about the most prudent course of action and what trade-offs should be made between appropriate regulation without stifling innovation and falling behind on the technology front. These debates have necessitated increased dialogue between legal and AI experts. In this seminar we will explore and acquaint ourselves with some of this informed debate and then zoom in on the issues regarding Intellectual Property w.r.t. the whole generative process. We will focus on the debate within the US as the debates are in English (!) and also many of the big technology firms driving the progress in generative AI are based there and thus what happens in the US will have an outsized global influence.

Background Overview

A great place to start is the recent workshop at ICML 2023 - Generative AI + Law (GenLaw) ’23 Links to an external site.. Everyone should (skim-) read this paper resulting from the workshop as it lays the foundation for many relevant issues:

  • Report of the 1st Workshop on Generative AI and Law Download Report of the 1st Workshop on Generative AI and Law  by A. Feder Cooper, Katherine Lee, .......

 

Copyright Issues

The US Copyright Office currently has a request out for comments on copyright and generative AI (Artificial Intelligence Study Links to an external site.). The document with the official call is available here: Notice of inquiry and request for comments Links to an external site..  The quote below taken from this document highlights their need to find concrete solutions to the issues raised by generative AI:

"Over the past several years, the Office has begun to receive applications to register works containing AI-generated material, some of which name AI systems as an author or co-author.4 At the same time, copyright owners have brought infringement claims against AI companies based on the training process for, and outputs derived from, generative AI systems.5 As concerns and uncertainties mount, Congress and the Copyright Office have been contacted by many stakeholders with diverse views. The Office has publicly announced a broad initiative earlier this year to explore these issues. This Notice is part of that initiative and builds on the Office’s research, expertise, and prior work, as well as information that stakeholders have provided to the Office."

Much of the document is a series of grouped questions covering in very specific detail the potential issues raised by generative AI.

From this page one can access Submitted comments Links to an external site.. Comments have been submitted by the relevant big companies, academics,  artists,  representative unions and individuals. Examples of submitted comments are

  • Talkin’ ‘Bout AI Generation: Copyright and the Generative-AI Supply Chain Download Talkin’ ‘Bout AI Generation: Copyright and the Generative-AI Supply Chain  by Katherine Lee, A. Feder Cooper and  James Grimmelmann
  • AI Art and its Impact on Artists Download AI Art and its Impact on Artists   by Harry Jiang, et al
  • The Verge has mined the submitted comments section for the comments submitted by the big players in the field and written an article on their thoughts on "copyright" payments:  AI companies have all kinds of arguments against paying for copyrighted content Links to an external site.

Most of the documents mentioned cite this work so it is probably worth skimming it.

  • Extracting Training Data from Diffusion Models Download Extracting Training Data from Diffusion Models by Nicholas Carlini et al USENIX Security Symposium, 2023

What you have to do for the seminar/assignment:

  • Read
    • Report of the 1st Workshop on Generative AI and Law Download Report of the 1st Workshop on Generative AI and Law  by A. Feder Cooper, Katherine Lee, .......
    • Notice of inquiry and request for comments Links to an external site.  - Sections I to IV
    • AI Art and its Impact on Artists Download AI Art and its Impact on Artists  by Harry Jiang, et al
    • Talkin’ ’Bout AI Generation: Copyright and the Generative-AI Supply Chain (blog) Links to an external site.  (Please feel free to skim the AI sections with which you are already familiar)
    • Choose one of the major players, Google, Meta, etc, from the Verge article you can see which companies submitted responses. Find this company's submitted comment at Submitted comments Links to an external site.  and read it. You can find the document by using the company's name as a search term. 

  •  Write a ~1000 word essay, informed by the background reading, on the company's comment you downloaded and read and include the following information
    • Why you chose this company
    • Summarize the main points in the comment
    • Choose one "major" point they make and explain why you agree with it - if you don't agree with any of them then choose the one that is most "reasonable".
    • Choose one "major" point they make and explain why you don't agree with it  - if you agree with all of them then choose the one that is the most "unreasonable/controversial".
      I don't mind if you use "gpt" to help you with the writing, but of course please be transparent about using it and check the text created is accurate!

  • At the seminar come ready to discuss the topic and the your essay. Please remember the seminar leaders are not experts in this area so part of the seminar to help inform them about the background readings and the company comments!
0
Please include a description
Additional comments:
Rating max score to > Pts
Please include a rating title

Rubric

Find rubric
Please include a title
Find a rubric
Title
You've already rated students with this rubric. Any major changes could affect their assessment results.
 
 
 
 
 
 
 
     
Can't change a rubric once you've started using it.  
Title
Criteria Ratings Pts
This criterion is linked to a learning outcome Description of criterion
threshold: 5 pts
Edit criterion description Delete criterion row
5 to >0 Pts Full marks blank
0 to >0 Pts No marks blank_2
This area will be used by the assessor to leave comments related to this criterion.
pts
  / 5 pts
--
Additional comments
This criterion is linked to a learning outcome Description of criterion
threshold: 5 pts
Edit criterion description Delete criterion row
5 to >0 Pts Full marks blank
0 to >0 Pts No marks blank_2
This area will be used by the assessor to leave comments related to this criterion.
pts
  / 5 pts
--
Additional comments
Total points: 5 out of 5