A growing community of online music fans is using Google Assistant's song recognition abilities to identify previously undiscovered samples buried in the music of artists like Daft Punk, Mobb Deep and Oneohtrix Point Never.
By running audio directly from their computers into the AI-powered song recognition technology used by the mobile virtual assistant software Google Assistant, sample hunters have been able to discover previously unknown samples, some even shorter than a second in length.
Members of the Discord community have identified samples in songs such as Mobb Deep's Hell On Earth (Front Lines), Quasimoto's Green Power, Modjo's Music Takes Me Back, Oneohtrix Point Never's Nassau and Daft Punk's Face to Face.
After making their discoveries, the sample spotters upload their finds to WhoSampled, an online database that catalogues samples used in contemporary music.
Google Assistant uses similar technology to song recognition service Shazam, but according to sample hunters, its accuracy is far superior.
"Google Assistant can even detect samples less than a second long, and is usually able to detect samples that have been chopped or time-stretched," community member DJPasta tells Tracklib.
Enterprising sample hunters have developed advanced techniques that help Google Assistant in recognising shorter samples.
"If the sample you are trying to find is part of a much longer drawn-out chord or texture, you can time-stretch or crossfade a loop to make the AI think it's longer. You can also repitch the sample to form a chord progression if you guess it correctly."
While this kind of technology has become a useful tool for sample spotters, some have voiced concerns that it could reveal the identity of previously unrecognisable uncleared samples in the work of independent artists.
"Can’t wait for a major label, RIAA, or whoever to buy this tech and start suing everybody, including me," writes sound artist Jake Muir on Twitter.