Nathan ClareGoogle’s AI, Google Assistant is a virtual assistant software based on artificial intelligence and has been an integral tool in uncovering hidden samples that were previously unknown to the public.
While advancements in the world of Artificial Intelligence now seem like the norm over the last year or so (primarily with ChatGPT or Text To Image’s Deepai), there had also been some discussion about how AI can and does interact with music. Initially, there was talk regarding AI being able to produce music entirely (and while experimentation is still ongoing), there has been recent advancements in how AI identifies samples that were unknown for as long as 20 years coming to the fore, thanks to Google’s AI, Google Assistant.
When people thought of Artificial Intelligence in the past, it is likely that they would have thought of sci-fi, or movies such as Ex Machina that depict AI as an antagonist against humanity, rather than something that could prove to be so useful on such a basic level for people all across the globe providing they have an internet connection and are aware of its existence. The growth of AI since it was released for testing last year has been truly unprecedented, and while dystopian feelings are more than valid it is impossible to deny the potential such a tool possesses in the right hands.
The growth of AI has been so rapid, seeing as it took ChatGPT as few as 5 days to reach 1 million users which is simply unheard of. We have now been introduced to software that can explain almost anything in a clear, concise, and easy-to-understand manner. As well as newer AI that can create AI art by inputting text into a search engine (example below). Of course, this can act as a massive scare to many who find themselves working in industries that may look to utilise AI to automate and cut costs in aspects of certain jobs, it does seem there is reason to maintain a level of cautious optimism with such a tool though despite such vast developments.
Looking at more recent advancements within the world of music and AI, there has been new ground made with being able to identify samples used by producers on numerous records that were previously unknown to the public by Google Assistant. While this can come as an exciting relief to many, it can also ruin the fun or the ‘grind’ that is involved with sample hunting for some. Possessing such a tool in many ways could feel similar to knowing cheat codes for a game for instance, and while it is fun at the beginning, you soon realise that the novelty can wear off.
A fascinating example of Google Assistant’s ability to uncover samples was utilised by DJPasta, who is an active member of the Sample Hunting community. They uncovered a method that would allow audio to run directly from their PC into Google Assistant with another software called Bluestacks. They were curious to trial some Todd Edwards samples and then shorter ones too like Carrie Lucas’ ‘Sometimes a Love Goes Wrong’ which led to the discovery of even more material that was used in Daft Punk’s timeless track Face To Face. There have since been more sample discoveries that include Mobb Deep, Madlib, Nujabes, Modjo, and more.
Another issue that could potentially erupt for a lot of artists and their work is that they could now find themselves in legal battles if they have profited from compositions utilising samples that might not have been identified or cleared in the past, though it might have been unknown to the public up until now. Although Youtube does use a relatively competent content ID software that assesses productions to ensure that it is not uploading copyrighted material that might not have been cleared, the capability of Google Assistant in this instance is far greater than that of the previous software used by Youtube. Meaning that there were still numerous tracks bypassing the software, providing the sample used wasn’t a blatant copy with very little done to it when used by another artist, which is good since no one would like for their art to be an essential carbon copy of somebody else’s work while it is described as a sample with very little work done to it in the first place.
Though, if a sample has been edited to the point where it is unrecognisable by humans, it now has a much higher chance of being flagged by AI. In legal terms, such a piece of software would be openly welcomed, but in terms of artists and individuals who may be advocates for freedom of expression through the medium of music production (and inherently sampling), this could cause massive complications, as so far as becoming much more limited in terms of how the material is sampled, edited and re-used by artists for new projects.
While there should be a significant level of security and protection in place for artists who lose out significantly from others due to people profiting massively off of their copyrighted material, a degree of discretion should be advised to ensure that it doesn’t restrict the capabilities of sampling past the point of no return. As that could be seen as the greatest overall loss and could have a massive cultural and societal knock-on effect as a result.
Out of curiosity, I asked ChatGPT ‘Could artistic expression by means of music production and sampling be at risk due to new AI technology that is available which can discover samples that humans were unable to find previously?’, this was its response. As you can see its response is clever and thought out. Though it can be said that it doesn’t take into account the legal, societal, and cultural complications it could cause.
Experiment with ChatGPT here.