IMHO these are word games with the effect of distorting the meaning of Copyright and Fair Use, and likely with the intent to drive an agenda. There is nothing scientific about the analysis; case law is used in support the theory but rarely to challenge the findings.
@alexjc
-
Copyright Exception 17 USC 117 Unlikely in AI Legal Battle
By
–
The basis for this is discussed in the technical section above. It's an argument that has also been made before already and contested, so there's not much new here either. 17 USC 117 is raised as a © exception, but frankly that's such a stretch there's no chance it'll fly.
-
AI Training Fair Use Argument and Copyright Infringement Liability
By
–
On the U.S. side, the argument is a common one that's been made many times before:
"Training is Fair Use." Reproductions of protected works *are* made during training, so AI companies could be liable for infringement, but the argument goes, that's Fair Use. -
German Legal Compliance Issues with LAION AI Training Data
By
–
Further, compliance under German Law is subject to a variety of details that haven't been taken into account — in particular in the case of LAION (who funded the work, was opt-out respected, were files deleted). Thus those conclusions about TDM in Europe should be thrown out.
-
AI Hallucination in German Law Analysis Reveals Understanding Gaps
By
–
The analysis of German Law was problematic from the start as the wrong translation was cited. It was a hallucination taken from a third-party blog rather than the original thing. That was fixed, but shows the level of understanding…
-
Weak Arguments Require Opposing Viewpoint Response
By
–
The paper has stimulated some fruitful discussions, but it does require the opposing viewpoint to write a paper because the arguments are not very strong IMHO and should not be left to stand unchallenged.
-
Legal Framework for Text and Data Mining in AI
By
–
[🅲] Legal Problems The paper focuses on the U.S. jurisdiction; however it digs into German Law in order to find legal support for TDM since there is no such concept in U.S. Copyright Law.
-
Categorical Assertions in AI Documentation Risk Misinterpretation
By
–
The major objections I have with it is that the body itself contains many categorical assertions, which are defended by the fact the conclusion is reasonable. Unfortunately, this means people will quote the details to support their own opinions, thus causing a downward spiral.
-
Authoritative Controversial Content Presentation Ethics Problem
By
–
If it was a less opinionated tutorial, or an opinion piece not presented as correct, this would be fine, but the middle ground is a 60+ page paper that looks authoritative and yet it's controversial by design and everything can be disagreed with.
-
Legal Protection for AI Companies: Training Data and Output Defense
By
–
The most value for lawyers and their lawfirms is in defending AI companies at the training stage 1) inputs, and then secondarily, 3) outputs. If the rights for model weights 2) can't be secured, then fine that's an acceptable tradeoff because 1) is critical.