Classical Music

Guest blog: Composer Thomas Hewitt-Jones on the threat AI poses to music


I am generally an optimistic person and tend to embrace new technologies. Innovation and change are arguably what shape society, and in many respects, we live in the most exciting era in history. Apart from dodgy MPs, vast inequality, the housing crisis, and the energy crisis. Oh…and Brexit. Anyway apart from all that, we live in an era where well-being at school is appreciated as much as grade A*s, and the young people of tomorrow know to embrace new technologies and take risks, as they know they are financially screwed anyway.

However, I cannot put this more strongly: AI poses the greatest threat to civilisation we have yet seen.

As a professional composer fast approaching my 40s, I’ve recently taken stock amid a busy stream of commissions over the last couple of years. I know I believe firmly (as do many of my contemporaries) that the value of human endeavour of any kind should not be free, and hard graft (of whatever nature) should be respected and valued.

But in thinking this all through, I have to go right back to the essential question: what is the purpose of art? For me, human creativity is the very essence of identity. The reason for being. My own artistic journey is about creating order from inherent disorder, where an individual vision (in whatever genre) has the potential to make its unique and lasting mark. It’s all about process, and I practically live by the Stephen Sondheim idea that “content dictates form”.

However, I do know that others see the pursuit of creativity through a more corporate lens. There are frequently up to 7 writers on a pop song, sometimes as many as 9. Film scores often have whole teams of additional writers, and orchestrators are usually employed for reasons of speed (even though their job is often more of a MIDI/audio transcriber and rearranger). Conversely, classical composers normally work on their own, and use a combination of inspiration and hard graft to craft an emotional narrative of horizontal lines and vertical textures, and I’ve always sat slightly on the more ‘commercial-but-aiming-for-integrity’ end of being a melodic classical and commercial composer.

When human ‘ghostwriters’ come into the equation in commercial music, I can sort of understand. But the really interesting stuff – the game-changing pinnacle of creativity – is always aware of the limitations of the human condition; it isn’t a coincidence that some of the finest human expression has been on down days or late in life, when staring mortality in the face. Documenting our journey through life – warts and all – is something that is hardwired: it’s a human compulsion, and those of us whose work represents personal fulfilment, but also reaches an audience, should consider ourselves extremely lucky.

“The ease of access to AI-generated content may decrease motivation for humans to create.”

cHATgpt

The current crisis on our hands (I do not use that word lightly) is that AI machine learning is about to get many, many times cleverer than it is already. Google’s foray into AI music is pretty crap so far, but it’s advancing incredibly fast and will surely replace low-level commercial music within a couple of years.

A good thing, one might say. But despite the lack of universal free music education (which we all dream of), audiences are smart. People can tell instantly when music has artistic integrity if they aren’t deliberately patronised by what I call ‘lowest common denominator music’ and are given a choice. Most sentient beings are open to art, and art can enrich them.

I should probably wind back a few paragraphs and unpick applications where AI can potentially really help humanity – and I asked ChatGPT to outline these for you here. This is what it said:

  1. Improving healthcare (e.g. diagnosis, personalized treatment plans)
  2. Enhancing education (e.g. personalized learning, grading)
  3. Driving scientific discoveries (e.g. drug development, climate modeling)
  4. Automating mundane tasks (e.g. data entry, customer service)
  5. Reducing poverty and inequality (e.g. creating jobs, improving access to services)
  6. Improving public safety (e.g. disaster response, crime prevention)
  7. Enhancing accessibility (e.g. speech recognition, language translation)
  8. Environmental protection (e.g. reducing waste, conserving resources).

However, unchartered AI let loose on the creative industries is incredibly scary for multiple reasons. Large AI companies profess to make the technologies ‘open’ but what they are really doing is writing code that steals from creative people. That’s what it does – it steals. And that’s not OK. Governments put ‘copyright exemptions’ in place to try and support these new and emerging technologies, but everyone is quickly realising that this will not work. A world where any new creative endeavour is immediately ‘harvested’ is not acceptable – not only in terms of financial loss to the creator but also in terms of the existential threat to humanity and people’s wellbeing. I have it on good authority from friends who work in the AI sector that technically speaking, a system of remuneration to the original creators of any ‘harvested’ work is entirely possible, but that governments need to make appropriate laws immediately, otherwise, this will be skirted over and not come to fruition.

Otherwise, we need to have the option to ‘opt out’ of data harvesting. I’ll certainly be writing that into all of my music contracts from now. There is a UK government committee about to investigate the ethics of AI, and in the US the White House an ‘AI Bill of Rights Blueprint’ has just issued, but more needs to be done globally specifically regarding the protection of copyright material else the very essence of human creativity is at stake. AI needs to be properly legislated by governments, and it needs to happen fast. 

I would like to think that developers working on AI and machine learning have good intentions at heart, and are all trying their best to keep the modern world fresh and exciting. The problems arise when technologies outpace governments and legal systems to the point that it becomes almost impossible to put the genie back in the bottle.

I’ll give the last word to ChatGPT.

I asked it to outline the threat AI poses to the value of copyright. It responded:

Outline the potential threat AI poses to the value of copyright.

  1. Ownership disputes: AI-generated content may raise questions about who holds the copyright to the work, the creator or the AI program itself.
  2. Difficulty in enforcement: AI-generated content can be easily copied and distributed, making it challenging to enforce copyright laws.
  3. Lack of originality: AI-generated content may lack the originality and creativity required for copyright protection.
  4. Rapid dissemination: AI can generate and distribute content at an unprecedented speed, making it difficult to control the distribution of copyrighted material.
  5. Unclear legal framework: The existing copyright laws may not adequately address the unique challenges posed by AI-generated content.
  6. Potential harm to human creators: AI-generated content may decrease the demand for human-created content and harm the livelihoods of creators.

And then I asked it: 

Outline the potential problems AI poses to the value of human creativity.

It said:

  1. Job displacement: AI can automate tasks previously performed by humans, potentially reducing the need for certain creative jobs.
  2. Decreased uniqueness: AI-generated content may lack the distinctiveness and originality of human-created content.
  3. Bias and lack of empathy: AI may perpetuate existing biases and lack the capacity to understand the nuances and emotional aspects of human experiences.
  4. Overreliance on AI: Overreliance on AI may stifle human imagination and problem-solving skills.
  5. Intellectual property concerns: Issues surrounding the ownership and distribution of AI-generated creative works are yet to be fully addressed.
  6. Decreased motivation: The ease of access to AI-generated content may decrease motivation for humans to create.

In conclusion, while AI has revolutionised the way we live and create, it also poses a threat to the value of human creativity and the protection of copyright. It is important to address these challenges and find ways to balance the benefits of AI with the protection of human creativity and the value of copyright.

That last paragraph was also written by AI.

So I would like to think that it agrees with me.

See the Blueprint for an AI Bill of Rights Blueprint for an AI Bill of Rights | OSTP | The White House

MusicLM – Google Research’s foray into AI music generation from text prompts  MusicLM (google-research.github.io)



Source link

Related posts

Classical home listening: Haydn with Harry Christophers, and a new Welsh-language opera | Classical music

Mary McCartney

Jane Antonia Cornish, Sierra, Vicky Chow

Mary McCartney

New York’s Shed shrinks British chief – SlippediscSlippedisc

Mary McCartney

Leave a Comment