Research is inherently different to commercial exploitation. There are often explicit clauses in laws & regulations allowing that. Don't conflate the two. It's a slippery slope fallacy.
@alexjc
-
Fair Use and Corporate Risk in AI Training Legal Debate
By
–
The fact companies claim Fair Use is only necessary if it's prima facie infringement—which legal scholars increasingly agree (see Workshop from yesterday). Given this, debating what's "legal" is weird because Fair Use is always case-by-case and corps intentionally took the risk.
-
Copyright and AI Training: Public Domain Content Misuse
By
–
Interesting you're defending her because that's the same situation that creators find themselves in against MidJourney. But, unlike those being infringed by MJ, she had placed her content in the public domain.
-
Companies Building AI Models Legally: Business Decision Not Inevitability
By
–
Well, if you agree it's a business decision then we've already corrected the "inevitable" part of your original statement — which was my intent. I don't understand your framing of the question. Companies large and small can build useful models legally.
-
Adobe Firefly Legal and Ethical Concerns Beyond Training Data
By
–
Of course, Adobe Stock also has many questions surrounding its legality and it should be scrutinized. But the fact that's the topic for Firefly is a step well beyond training on infringing content, or worse illegal content. Again, it's a business decision.
-
Licensed Content in Diffusion Models: Business Decision Not Inevitability
By
–
The fact that there are diffusion models with properly licensed content suggests it's not "inevitable", but rather a business decision.
-
ML Models Cannot Distinguish Copyrighted Patterns Judges Say
By
–
Ultimately, the burden of proof is on them? He's wrong in that there is no mechanism built-in to distinguish Copyrighted patterns from non-Copyrighted ones. Even judges struggle to find the difference, and somehow we're supposed to believe ML engineers solved that accidentally?
-
Copyright Fair Use Debate in AI Training and USCO Submission
By
–
The USCO submission he has his name on argues it's non-expressive use, which he now seemingly disagrees with by saying it falls under Copyright but should be considered Fair Use.
-
Internet Data Scraping and Copyright Infringement in AI
By
–
❝ Just grabbing information off the internet is *definitely* copyright infringement. ❞
-
Strategic Lobbying Behind AI Power: Beyond Emotional Reactions
By
–
To portray the powers at play as reactive and emotion-driven is misleading, as the lobbying force behind them is IMHO strategic and calculating — almost devoid of emotion to the point of psychopathy.