✰ = Source Link for citation, click to see where tidbits were taken from.
✰ Jinriki is a common practice in NicoNicoDouga, and the western UTAU community, and involves re-editing an existing singing voice (especially a singer, game character, or anime character), extracting tones as WAV files and reassembling them. It is most commonly seen in UTAU, with hundreds of Jinriki voicebanks having been made. UTAU was originally created to assist this process using concatenative synthesis. UTAU is able to use WAV files provided by the user, so that a singing voice can be synthesized by introducing song lyrics and melody.
Though, as of recent, there has been a lot of Jinriki voicebanks utilizing AI technology. More notably, the program DIFF-SVC. This has brought in many controversies and legal issues surrounding the Jinriki community. Even causing DIFF-SVC's community and resources to become privatized to avoid Jinriki voicebanks.
✰✰ Jinriki UTAUloids almost explicitly contradict UTAU's Terms Of Service in regards to the legal sourcing of voicebanks, as well as general copyright and consent rules when it comes to others' voices used without their permission. Of course, there are Jinriki that do have the voicers consent, thus making them completely legal to be distributed publicly. Ameya [the creator of UTAU] decided to stop creating Jinriki in general since it goes against copyright. Distributing, sharing, and using Jinriki for monetary use has often resulted in being blacklisted from many UTAU spaces.
Though there's a bit of a legal loophole when it comes to UTAU Jinriki. As long as you don't publicly distribute your Jinriki UTAU, and use it for personal use, there's no real way to be blacklisted. Of course, the question of whether or not it's ok to recreate a voice without the voicer's permission can be argued. However, legally you can't really be blacklisted if nothing's ever shared. This loophole is often taken advantage of by people who create illegal Vocaloid ports. Also, the existence of parody laws may protect certain Jinriki, especially those who have different names/designs than their voicers or those who don't sound indistinguishable from their voicers.
✰ Jinriki legality is not an UTAU-exclusive issue, especially as of recently. Videos of a DIFF-SVC AI voicebank made using Ariana Grande's voice have gone viral, and even gotten the attention of several news outlets. The voicebank was almost unable to be differentiated from the source, causing the question of copyright infringement and consent to rise. The fear that AI celebrity voicebanks could be used for defamatory or otherwise infringing purposes was heavily discussed. This caused DIFF-SVC's public resources to become privatized. DIFF-SVC is different from UTAU, in the sense that it's even harder to distinguish from the actual voicer itself. As in UTAU, the voices are very obviously robotic-sounding and thus could technically be pronounced parody work. It is never suggested to use DIFF-SVC for Jinriki purposes without the voicer's consent, as it's illegal.
Jinriki voicebanks are very different from ported voicebanks for a multitude of reasons. Though it's possible for ports to be Jinriki and vice versa, both terms are very different. With Jinriki being an umbrella term for voicebanks made from external voicers, as previously stated. Ports, however, are a lot more easily defined. For example, if you were to take a bunch of vocal samples from Hatsune Miku's Vocaloid voicebank and edit them to be compatible with UTAU, that would be a port. Also, technically a Jinriki of Hatsune Miku's voice. However, it's a port from one synthesizer to another. There have been countless authorized and unauthorized ports, and the legality is very grey, but it's an extremely common practice in the community.
Making Jinriki is extremely easy with current-day technology. In fact, all you'd really need are samples of the voice you're trying to recreate (in the WAV format), audio editing software (Audacity is commonly used), and immense patience. It's suggested that the samples you use don't have any background noise, as it's harder to edit. The most common form of Jinriki creation is categorized by taking various WAV samples and opening them in editing software to trim out certain sounds to create syllables. For example, if you have a voice clip that is singing the word "hello", you'd cut it into two different sample files ("he" and "lo").
Here are some tutorials that are helpful in learning the ways of the Jinriki.
Here are some of the most popular examples of Jinriki.