For Assignment 1– Composing Memoirs with Sound, I intend to create a compilation of all of the artificial sounds that I hear on a daily basis. This will demonstrate how the sounds created by technological advances changes the way in which our auditory senses perceive the world around us. The earliest humans on earth lived in a world void of so many of the artificial sounds that are so commonplace to us today. The sounds of nature would be so much more prevalent in those days – from the sounds of birds chirping to the the wind blowing, I’m sure the world sounded vastly different than it does to us in today’s age. Sounds such as airplanes flying by overhead, the roar of traffic on nearby roads and highway systems, the hum of a computer, the buzz of the air conditioner or heating unit – these are all sounds that have become so commonplace that we have essentially learned to tune them out. But by focusing solely on these intrusive sounds, this project will examine just how prevalent these sounds are to the world in which we live.
So in what ways has A.I. affected the world of sound? I think we are all familiar with A.I. generated speech applications such as Siri or voice-enabled GPS navigation systems and software. A brief history of voice assistance technology can be seen here
Recent developments in the advancement of A.I. and sound would blow your mind. MIT researchers developed algorithmic technology capable of replicating sound. The technology takes video recorded of an object (in the demonstration video a drumstick was used) striking other objects and surfaces, with the end result predicting with uncanny accuracy the sound effect that correlates with the video recording. This technology has numerous real world applications and could be used as simply as more realistic sound effects for movies or could be used to give robots a better understanding of objects in their surroundings.
Another astounding development is the recent unveiling of a technology called Tacotron 2 produced by google. This technology enables near indistinguishable computer-generated voice from that of an actual human. The link gives examples of sentences read by both the A.I. voice and an actual human – can you tell which is which? I sure couldn’t!
A.I. also has its malicious side when it comes to sound, as evidenced by computer scientists at USC, Berkeley who developed an A.I. -based attack that exploits vulnerabilities in speech-to-text systems such as the DeepSpeech system. To quote the YouTube video produced by the scientists:
The attack works by playing specifically crafted obfuscated commands that are easily recognized by the targeted device but are difficult to understand by human listeners.
This technology poses a serious threat to the ever increasing world where voice-controlled devices are rapidly increasing in popularity. It’s not hard to imagine this technology being used to implement attacks against devices such as phones, computers, cars, or TVs which would effectively render the systems useless.
These are just a few small examples of the ways in which A.I. is permeating the world of sound. Don’t get too comfortable with the way the world around us sounds right now – because things are about to become much crazier! (And more digitized).
A.I. – we are all familiar with the term. It is the future, after all, and we are increasingly seeing our own lives becoming immersed in its world after only reading about it novels or seeing it on television shows and movies. From self-driving cars to Google’s DeepMind initiative to Amazon Alexa, A.I. is all around us. This blog seeks to delve into the inner workings of A.I. and how it relates to sound. I further hope to address this concept – is A.I. REALLY artificial intelligence, or is is as close to actual intelligence that we can achieve? Human reasoning and analytical ability is often clouded by emotion and ethical concerns, whereas A.I. is a principal guided strictly by logic, by ones and zeros, and by yes or no commands and inputs. I feel this gives machines the edge over humans. But I will explore this particular topic in further in later posts. For now, I will examine how A.I. relates to sound.