Referencing

 Why is my two pheno books without a good referencing, not a footnote nor endnote, but AI's own word?

Because it told me so that it is its own synthesized wordings. If it is a direct quote of an author, it told me it will tell me so (but might as well hide the reference to it). And because we are chatting to each other he just as if refer to it in passing. Maybe if it is a scholarly AI then it will always have it's sources all referenced. For example, while a Google AI Mode will always have it's references with it because it's searching the whole online content, Gemini wouldn't.

If Sony and UMG will win against Suno and Udio, the court rulings will trickle down into a strict referencing of the sources of LLM AI as well.

If music is just for feeling, then this book two has started from a wrong footing that HM both has feeling and meaning. But it seems not. Then HM is also about meaning. But they are fighting for similar sounding which by itself is meaningless without words attached to it. Lyrics would be easier to spot on as similar to something already sang in the past with exactitude. Plagiarism is well avoided I think by an LLM, and it's not also the battle in court all about.

There is no common knowledge it seems as well in music, pure sound like a base drum maybe, though accurately it is called royalty-free sound like what BandLab provides for its user. Nevertheless, you still have to mention where did that sound came from for one's safety and clearance from suit, unlike the common knowledge sentence the earth revolves around the sun, which needs no referencing anymore.

I just would like to mention that a lo-fi music I found in YouTube was meant for deep focus music I subjectively interpreted differently as a music coming from my broken heart. But generally we've culturally crafted what sounds are happy, what are sad. A music even without a lyrics has some inherent meaning in it. It might not be general, but we don't usually ecstatically dance to a mournful music.

In contrast, we don't reference music, unless of course Sony or UMG tells the court this Suno or Udio song got its similar sounding song from this HM. I just don't know how similar they need to be for the court to declare a song infringing.

...The Songs: "My Sweet Lord" (George Harrison, 1970) vs. "He's So Fine" (The Chiffons, 1963)

This is likely the foundational case you heard about. It is famous because it established a massive legal precedent: subconscious plagiarism...

...Marvin Gaye won... This decision terrified songwriters. Previously, you could only infringe on a copyright by copying specific lyrics or melodic notes. The Blurred Lines decision meant you could be sued just for replicating the "groove" or "feel" of an older track. It is the reason why artists today proactively give writing ncredits to older musicians just for safety. Music Copyright Lawsuits

 Well someone can argue I did not intend to be born having the same look as you are. But Gemini 3.5 Flash confirmed this sentiment of Milan Kundera that similar looking people, even identical twins, don't look exactly the same via facial recognition biometrics we have nowadays. Yilong Ma won't be sued by Elon Musk just for having a similar looking face. It's not the case for a song though.

Do you agree this music legal disputes will later face a dead end, unless of course two songs are actually too identical already? The meaning (lyrics) behind are not seemingly that important to the case. As a music composer myself, being afraid to be sued later kills any kind of inspiration. The human tendency to create suffers much. Perhaps an algorithm to detect if my song infringes none, must be developed and promulgated by copyright organizations as an international standard not just third party like Cyanite.ai and Audd.io. Composing music becomes a get-ahead-and-own-everything scenario, not an "I compose because this is meaningful to me." Just don't say to me what you also have in mind that gen AI music will own everything because it is already all in its storage or even just in its inception...

Like attracts like. They tend toward the same direction. Today is 31st of May when luwa is recited here in the Philippines this Flores de Mayo. I just felt a kind of awe that ladies know the Blessed Mother more than us boys. A sad song will tend to sound sad. I think we just have to respect emotions and not just box it into "it's infringing something." In the end, it is not just an emotion or a music or a sound. Literature says poetry are for speaking when prose fall short of what the human heart wants to say from its abundance. And a song, be reminded, is not just a poem. It captures a musical reality that words and logic could not express anymore.

Even a quantum computer computing all possible chess combination nonstop would not finish before earth is consumed by the sun. A Ruy Lopez opening alone which looks similarly the same (otherwise they wouldn't be categorized in that opening) has over 200 named variations. But that's not the case for music nowadays. By copyright issue, they cannot stand as a variation! As if, you need to create the Sicilian opening just to make it different from Ruy Lopez. And the speed at which AM is created has more than doubled the speed of music creation overall, which speeds up the limit of how many songs are still creatable given the present state of affairs and laws.

I just don't want to imagine that time will come, I need to compose a very different kind of music very far from what is inspiring me or not sounding like a music anymore, just for copyright office to accept it as not infringing on any existing song already. 

That sounds like the end of copyrighting as we've envisioned. The nightingale birds will celebrate they can sing to their heart's content without being sued consequently. And even the street musicians can be subjected for collections if looking poor hides an auditable higher than living wage for a week. Music will just be one time performance only since recording and uploading of it even in private groups can be spied as playing similar sounding songs. And just like Einstein regretting urging US to look into nuclear  weapons, Thomas Edison could have had some regrets as well how the phonograph ended musical creativity of human beings that only machine can do with exactitude how to bypass the copyright office.

It's eerie to imagine police raiding a hidden live music venue because they are playing songs of private individuals, not published or cleared by copyright office to be played as they are 10% to 20% similar sounding to other older songs.

Given the current state of affairs and laws, mathematicians should work with copyright office when this feared consequences is expected to happen, and what can be done to fully prevent it. Be reminded that copyright offices aren't primarily protecting the rights of artists and authors, but the encouragement of the act of creating. If just 1% similarity of all combined elements (given the algorithm or mathematical formula is the best one) becomes a copyright infringements for songs, then we are already nearing the end of song creation. Or will it still sound musical?