Artificial Intelligence (AI) AI and Music Education
Zachary Keita, NAfME Advocacy and Public Policy Communications Manager Republished with permission of the National Association for Music Education (NAfME).
The dawn of the 21st century brought with it a wave of technologi-
cal advancement such as mobile phones, the Internet and globalized communication. All these tools work together to make information more accessible than ever before in human history. Countless in- dustries have benefited from this newfound availability of informa- tion, including the education sector. The Internet provides students with a wealth of information from which to draw, while teachers can access a breadth and depth of resources that dwarf what was available to their 20th
century counterparts. These technological
advances have fundamentally changed the classroom dynamic and will continue to do so as long as technological evolution persists at such a rapid pace.
The latest breakthrough advancement in technology, Artificial Intelligence (AI), has the potential to upend the education system as we know it. The U.S. Department of Education (ED) defines AI as “automation based on associations,” while IBM defines AI as “a field, which combines computer science and robust datasets, to enable problem-solving.” The common thread between these definitions is the ability of AI to automate problem-solving based on associations drawn from robust data sets.
As new technology becomes integrated into everyday life, it’s important for educators to understand the potential implications for their students and their classrooms. Recently, AI has begun to make its way into the music space, lowering the barriers to entry for music creators even further by enabling high-quality production without the need for professional equipment. Currently, there are several AI-powered programs capable of generating instrumental soundtracks based on text prompts. AI in the music space also brings a host of new copyright concerns. Recently a U.S. federal judge ruled that “AI-generated art cannot be copyrighted,” not- ing that AI-created works lacked human authorship, which is a “requirement of copyright.”
As it concerns the music education classroom, AI has the potential to revolutionize how music is taught. Currently, there are several AI-powered programs that would enable students to create their own music with minimal input from educators. Examples of AI being used in the music education space are limited, but NAfME will continue to track policy and regulatory developments con- cerning AI in education, particularly around its potential use in the music classroom.
Types of AI
At this point in time, AI can be broken down into three catego- ries: Narrow or Weak AI, General or Strong AI, and Artificial Superintelligence. The two latter categories of AI are purely theoretical, as no such system has been developed. General or
30
Strong AI would possess human-like cognitive abilities and be able to apply knowledge across a wide range of topics, while an Artificial Superintelligence would possess cognitive abilities that surpass that of a human and be able to perform tasks that we currently consider impossible. In the next section, we will focus on the type of AI currently available and rising in prevalence, the Narrow or Weak AI.
Narrow or Weak AI
All currently available forms of AI fall into the Narrow or Weak category, defined as “AI designed to perform a specific task or a narrow range of tasks.” This type of AI has been available for over a decade and has become a tool used in everyday life. Examples of Narrow AI include virtual personal assistants (Siri, Alexa, etc.), customer service chatbots, autonomous vehicles and language translation programs. Besides autonomous vehicles all given examples of narrow AI have been integrated into daily life seamlessly. There has been no need for legislation heavily regu- lating the use of Siri or translation programs, as their use is rather limited. The more recent models of AI, such as ChatGPT, are capable of writing papers from simple prompts, solving complex math problems and much more.
The recent political fervor around AI comes as a result of the latest breakthrough in AI technology, Large Language Models (LLM), becoming widely available to the public. Rather than explain an LLM, I will demonstrate its utility.
The following section was generated using OpenAI’s ChatGPT, posing the question, “What is a Large Language Model?”
“A large language model refers to a type of artificial intelligence (AI) model that has been trained on vast amounts of text data to understand, generate, and manipulate human language. These models use advanced machine learning techniques, particularly in the field of natural language processing (NLP), to capture patterns, relationships, and semantics present in written text.
Large language models can perform a wide range of language- related tasks, such as text generation, translation, summarization, sentiment analysis, question answering, and more. They learn to generate coherent and contextually relevant text based on the patterns they have identified in the training data. These models are often used as the foundation for various AI applications that require understanding and interaction with human language.”
Effect on Education
A main point of contention in the debate around AI is its poten- tial effect on education. Specifically, the ability of LLMs like
April 2024
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62
