AI, babies, bathwater & the future
WARNING: No images, videos, explainers.
Whether in the modern age we have the attention spans to read longer form articles anymore is somewhat hard to know. But here goes…
There is some fascinating research coming out about emotions and decision making (executive function). Long story short: when a person is in the grip of an emotion, brain imaging shows that the areas of the brain responsible for thinking rationally, logic and objectivity is actually shut down by the emotional state. It can’t function again until the emotion networks are dialled down. ‘Amygdala hijack’ is in some ways a powerful animalistic survival program; in other ways a terribly detrimental flaw still in our design. Anyway, just something to consider whenever faced with extreme emotion and trying to talk rationally without confirmation bias: it’s not always our fault but might not be productive.
AI has been around arguably since the 1950’s and has been researched and worked on mostly in academic settings ever since. By the 90s is was termed ‘machine learning’, then later ‘algorithmic AI’. To today, the topic of debate being described as ‘generative AI’. As GPUs get more advanced it opens much more potential, and for some: more concern.
So, can AI can be just stopped and got rid of by public opinion? Maybe, yes anything is possible. But you always have to ask: something is possible; but is it likely?
We could get back to toothpaste and tubes, but really the main question here is whether that’s the best route to even try to attempt, and whether by looking to history we might be unaware of it repeating as it always has. ” This time it’s different”.
IS this actually different from all of human history? Or is this the same thing repeating over and over?
In 370 BC as writing became more widely used and taught, people in ancient Greece were fearful they would lose their memories. The outcome was that writing transformed education and civilisation. By the 1400s we start smashing up printing presses, largely driven by a fear cheap books would undermine the value of the big book, the Bible. Where of course it accelerated literacy and science. In the 1800s there was a movement against steam trains claiming that moving faster than 30mph would damage the human body. Electricity would cause Blindness. Radio would stop families talking. TV would destroy imagination. Crypto was going to collapse fiat money, now our gran’s have some in their pensions. And now here we are, with the latest new thing. Or is it entirely different? In the last few years, the timeline has sheared and we have started a new reality entirely…
There is no question all these technologies, AI included, can be a force for good or bad, evil, even. But if we had thrown the baby out with the bathwater at any of these stages of our history, what would have happened? A thought experiment.
In this instance: imagine you are a sample designer, maybe you have been in that industry for 30 years and AI comes along. What are you to think? Well on first thoughts it could be concerning. But this all comes down to how it’s envisaged. Next then, image you are a producer who collects samples and uses them. What are you to think? It might be reasonable to be concerned AI will change your experience.
The old format: the artist or sample designer creates a sound destined for a pack. When that gets downloaded it gets downloaded by multiple people, the same sound.
But what is this a possible new way that keeps the core principle while bringing something magical? What if that sound is transformed into 100 different re-versions and the artist is credited for all of them and paid for all of them? 100 different people get 100 different sounds. What does that mean?
Does that earn the artist more, or less for their natural intelligence? Does that cost the user more or less for their purchase? And does that circulate more or less diversity into the rivers of the music production world?
Stealing music (sampling) stealing music (torrents) stealing production samples (“sharing”) is that the same again as the ancient Greek issue, or entirely different to modern AI use? It all comes down to correct and proper crediting. How many people strongly against AI have had drives full of stolen music, TV shows and cracked software? Not a claim, but a possibility.
There was a period where it does indeed seem early AI models were built on ‘scraped’ data, possibly an ethical oversight, maybe done nefariously. But the real question is how things are shaped moving forwards? If AI is around for a 1000+ years, the first 5 years were perhaps a time it went through the messy period to figure out how it can be used, ethically, productively and in a way that just makes everything better, more fun and more forward thinking for everyone. It will take collaboration, frameworks, legislation and lots of communication, but without question it’s possible. If we can land a space rocket backwards, we can definitely get that done, and quickly.
So be clear, we are personally not going down the route of full model-based training and generative re-creation, for the reason we just feel it’s better to have the purely human-made sound to be the core, the place of origination, always will. But then using human-made AI tools, to modify and variate that human-made sound using human made synths is a place of extraordinary, jaw dropping, astonishing, potential. Provided downloads are legal, artists are paid just as they always were.
But that’s just what we think, what do you think?
Back to News