Panic At The AI Disco? AI music in 2023 exclusive

Leave a Comment / AI Music, AI-Generated Music, Copyright, EU Copyright, Music Industry, Policy, Regulation, UK Copyright, US Copyright / By

Interest in AI surges to an all time high in 2023

2023 was a pivotal year for both artificial intelligence (AI) development and the music industry. AI music has firmly established itself as a game changer, capable of composing, producing and performing new music, causing concern and unease for record labels, music publishers, artists and songwriters. Ghostwriter’s release of ‘Heart on My Sleeve’ in April 2023 ramped up critical conversations about AI, intellectual property and personality rights as artists found that they have little control over how their voices are used. OpenAI’s continued development of ChatGPT and the recent spate of controversial voice cloning song releases have brought AI music to the forefront of AI debates. One can’t help but wonder what this pivotal year could mean for the future of music and the role of artists in the digital age.

The apparent success of AI music in 2023 follows a year of huge investment in AI music and notable AI music company acquisitions in 2022. Warner Music Group joined a $14m funding round with generative AI music platform LifeScore, AI music generator Soundful raised $3.8m of seed funding, Apple acquired British startup AI Music, and Spotify acquired AI voice platform Sonantic. Several AI music platforms became publicly available (Infinite Album, Beatoven, Splice, SongStarter), there were genuine AI music releases, such as Benoît Carré’s album Melancholia under his Skygge allias, created using AI tools developed by Spotify, and wallyPDF’s computer-generated lo-fi single ‘Beach Dreams’. 2022 also saw Holly Herndon showcase her own voice cloning technology Holly+ and the launch of Riffusion, an AI model that trains on images of music called spectograms. Developments in law and regulation were also bubbling under the surface in the run up to 2023 as the European Commission presented the first regulatory sandbox on AI and a pivotal class action lawsuit was launched by lawyer Matthew Butterick against Microsoft, GitHub and OpenAI over copyright issues and the use of open-source code with restrictive licences. Meanwhile, Tencent Music in China announced that over 1000 songs with human mimicking AI vocals had been released and one of them had accrued over 100 million streams. If 2022 was a simmering volcano, 2023 was the eruption: AI music development in 2022 paved the way for groundbreaking advancements in 2023 that could change the music industry as we know it.

2023 was the year that over 1000 technology leaders and researchers called for AI developers to pause work on the most advanced systems, citing profound risks to society and humanity. 2023 was also the year that saw the UK host the first global summit on AI safety at Bletchley Park to action safe and responsible development of AI around the world. However, in a year marked by anxieties about AI’s broader impact, this post focuses on the developments of AI music throughout 2023 which raise serious ethical concerns regarding the impersonation of artists and the lack of authenticity in AI-generated music, an emerging industry which threatens to undermine creativity, a trait so intrinsically human. This article serves as your comprehensive guide to 2023’s AI music revolution, tracking each month’s technological advancements, industry responses, and evolving arguments for law and regulation.

What do you think the future of AI and music looks like? Comment below!

January 2023

AI music tech news

Law and regulation

  • Class action lawsuit launched against Stability AI, Midjourney and DeviantArt regarding the use of copyright protected images to train AI models. While specifically tackling copyright infringement of images, the legal outcomes of this case will have knock-on effects for all creative industries, including music.

February 2023

AI music tech news

  • BMusic AI launches following three years of development using blockchain to promote transparency and trustworthiness for music creation and distribution, and to streamline royalty payments (this press release suggests that users can claim copyright for the generated piece, this is not necessarily true).
  • Aimi launches a live YouTube channel to stream ‘ethical’ AI-generated music.

Music industry news

Law and regulation

  • Following huge industry backlash, the UK government presses pause on its decision to amend copyright law to allow a TDM exception for commercial use, which would allow AI companies to mine creative works for data to train AI models.

March 2023

Music industry news

  • Sir Lucian Grainge, managing director of Universal Music Group (UMG), calls for a new streaming payout model to crack down on algorithms that guide consumers to generic music that lack meaningful artistic context and lower-quality functional content generated by AI.
  • The Human Artistry Campaign launches, spearheaded by Universal Music Group (UMG) and backed by numerous music industry organisations which outlines seven core principles to protect human creators from the rise of generative AI.
  • AI tools take the front seat in the promotion of AI for music creators (Billboard Pro, Spiceworks, Audiocipher)
  • SPLY85 releases new album created using AI tools.

Law and regulation

  • The United States Copyright Office has its say on AI-assisted works and copyright: songs and other artistic works created with the assistance of artificial intelligence can sometimes be eligible for copyright registration, but only if the ultimate author remains a human being. “When a prompt is used to generate visual or musical works, the traditional elements of authorship are determined and executed by the technology – not the human user“.
  • The UK government publish a white paper titled A Pro-Innovation Approach to AI Regulation, which suggests that it might be reviving its previous decision to amend copyright law in favour of AI developers. This approach to regulation intends to drive growth and prosperity, increase public trust in AI, and strengthen the UK’s position as a global leader in AI. Regulation will be based on a flexible, principle-based framework to include: safety and security; transparency and explainability; fairness; accountability and governance; and contestability and redress. The government aims to use existing regulators to implement the regulatory framework. The UK government wants to implement a sandbox approach to regulation where businesses can trial products and services with consumers.

April 2023

AI music tech news

  • Splash launches text-to-music AI platform BeatBot and claims it to be the ‘Midjourney of music’, said to outperform Google’s MusicLM.
  • Beatly Music announces its intention to build a takedown-proof home for AI music.
  • Deadmau5 founded company Pixelnyx teams up with Beatport, mau5trap and Pioneer DJ to launch KORUS, musical AI companions. Pixelynx offesr artists and creators the ability to play, create and remix music from a global network of artists and labels while using officially licensed music.

Music industry news

  • The Human Artistry Campaign sees a membership spike.
  • Analysts suggest that AI may be an existential threat for major labels.
  • Music industry asks Apple Music and Spotify to block AI bots that scrape for data to train AI music models.
  • Following several attempts at Fake Drake tracks, including voicing tracks by Cardi B and Justin Bieber, a track released using fake Drake vocals to rap Ice Spice’s ‘Munch’ causes Drake to call it “the final straw“.
  • Fake Rhianna covers Beyoncé’s hit ‘Cuff It’.
  • Ghostwriter releases ‘Heart on My Sleeve‘ featuring fake vocals from Drake and The Weeknd, causing major waves within the music industry.
  • AIsis releases an album called The Lost Tapes inspired by the music of Oasis in their heyday. The album was written and recorded by British band Breezer, but AI was used to recreate Liam Gallagher’s voice to sing the lead vocals.
  • Grimes announces that she will allow AI music generators to use her voice without penalty with a 50% split of royalties.
  • Mainstream media and music industry press push the narrative that AI will transform the music industry (The Telegraph, Evening Standard, Why Now, Music Industry Blog, Hypebot, The Guardian).

Law and regulation

  • China proposes new rules to restrict AI development: tech firms will need to submit security assessments to the National Network Information Department before AI services can be used by consumers, to include AI-powered tech used to generate text, pictures, sounds, videos, code and other content based on algorithms, models and rules. AI service providers in China are required to ensure that users understand the content generated by AI should not be used to damage the image, reputation and other legitimate rights and interests of others.
  • Universal Music Group is successful in removing Ghostwriter’s ‘Heart on My Sleeve’ from Spotify and other streaming services. Despite using AI generated vocals of Drake and The Weeknd on the track, it was a copyright claim that saw the downfall of Ghostwriter’s track after racking up hundreds of thousands of streams and over 15 million views on TikTok. While the song itself was an original, human-created composition, Ghostwriter did use a “producer tag” in the song by a producer called Metro Boomin, which allowed UMG to issue a takedown notice.
  • April’s voice cloning phenomenon highlights the need for serious conversations about personality rights. Called ‘Rights of Publicity’ in the US, currently only 29 states recognise rights of publicity. Common law jurisdictions, like the UK, do not have a specific personality right, but false representation could be actionable under the law of passing off. Clarity is needed across territories.

May 2023

AI music tech news

  • Google releases MusicLM to the public. MusicLM can turn text descriptions into music and is available on Android or iOS. Google says it has worked with musicians in finding out how AI can empower the creative process.
  • Itoka, a company that offers music creation tools powered by AI, offers a decentralised music platform developed to enable data self-sovereignty and permanent music storage, alongside digital rights management. Itoka seeks to tokenise AI-generated music content on the blockchain so that creators can licence that content and receive compensation. (As mentioned previously, any assumption that copyright subsists in an AI-generated work and can be licensed and generate royalties is not strictly true).

Music industry news

  • Music consumers go crazy for voice cloning, sparking a new era of fandom.
  • A new deep fake track reincarnates John Lennon’s voice to duet with Paul McCartney on McCartney’s post-Beatles song ‘New’ and fans are beside themselves.
  • Grimes launches her AI project Elf.Tech in beta with CreateSafe which allows anyone to clone her voice for their own music. By mid-May, over 15,000 people had used it.
  • Paris-based company and parent of music distribution company TuneCore promises to block 100% AI-generated music from being uploaded to streaming services through its platforms.
  • Music streaming and sharing platform Audius, a decentralised music community, announce that artists and labels will be able to ‘opt-in’ using an ‘AI-Friendly’ button on their settings page, if they wish to allow artists to consent to interactions with AI-generated tracks uploaded by other artists and fans.
  • Spotify deletes tracks generated and uploaded by AI music platform Boomy after detecting stream manipulation. This follows a decision to stop publishing new releases from Boomy in order to review “potentially anomalous activity”. (Boomy tracks were reinstated later in the month).
  • Timbaland uses AI to release a collaborative track with The Notorious B.I.G., stating that he “always wanted to work with Big, and never got the chance to… until today”. Timbaland also announces that he has a plan to commercialise AI software in the future.
  • A scammer makes thousands of dollars selling leaked tracks he claimed to be by artist Frank Ocean, only the tracks were really generated by AI. The scammer reportedly made 13,000 Canadian dollars from the tracks created using a model trained on very high quality vocal snippets of Frank Ocean’s voice.
  • French streaming service Deezer annouces it is building a set of tools through its proprietary ‘Radar’ technology to detect AI-generated content on its platform. Deezer wants to develop a system for tagging music that has been created with generative AI, including tracks that use synthetic voices.
  • It’s announced that Paul McCartney might be using AI to create a new Beatles track. Fans speculate that it might be the song ‘Now and Then’.
  • Inspired by the success of the fake Drake track that went viral in April, Eclipse Nkasi creates an entire album in three days using AI. Nkasi wanted most if not every creative element of the album to be created by AI, so he used different tools to work on different parts of the songs.
  • Universal Music Group (UMG) signs a deal with AI music company Endel, describing it as “a strategic relationship to enable artists and labels to create soundscapes for daily activities like sleep, relaxation, and focus by harnessing the power of AI”.
  • Sony Music chairman Rob Stringer announces that Sony Music is focused on fighting against low-quality content (such as AI-generated music) and protecting their premium quality artists. Stringer said he’s optimistic about the ways in which AI can be used to police the unauthorised use of Sony’s copyright protected assets.

Law and regulation

  • Microsoft’s chief economist, Michael Schwarz tells attendees of the World Economic Growth Summit that “we shouldn’t regulate AI until we see some meaningful harm that is actually happening, not imaginary scenarios“.
  • The UK government’s Intellectual Property Office makes a landmark pact to improve streaming metadata. Signatories include representatives of major and independent record labels, publishers, creators, streaming services and trade bodies. The pact includes the creation of a working group between music business stakeholders to explore actions on remuneration for existing and future creators. This could have implications for AI-generated music in the future if AI labelling through metadata becomes a requirement.
  • Members of the European Commission call on tech giants to start labelling AI-generated content on a voluntary basis before it likely becomes law.
  • The EU’s AI Act takes a step closer to realisation of it becoming the world’s first set of rules on artificial intelligence. The European Parliament’s Internal Market Committee (ITRE) and the Civil Liberties Committee (LIBE) adopted the draft mandate with 84 votes in favour, 7 against and 12 abstentions. MEPs aim to ensure that AI systems are overseen by people, are safe, transparent, traceable, non-discriminatory, and environmentally friendly. They also want to have a uniform definition for AI designed to be technology-neutral, so that it can apply to the AI systems of today and tomorrow.
  • The UK’s Competition and Markets Authority launches an investigation into AI foundation models. The CMA will focus on the competition and consumer protection considerations of developing foundation models, AI systems that focus on training one model on a wide range of data and adapting it to many applications.
  • The US Senate Judiciary Committee meets to discuss AI regulation. Senator Marsha Blackburn questions OpenAI CEO Sam Altman over how songwriters and music artists should be compensated when their works are used by AI companies.
  • During a US House Judiciary IP Subcommittee hearing the same week, Congress stress that they don’t plan to make hasty decisions regarding the regulation of AI.
  • Leaders of the G7 nations call for the development and adoption of international technical standard for trustworthy AI, with Japanese representatives noting that “the governance of the digital economy should continue to be updated in line with… shared democratic values”.
  • OpenAI CEO Sam Altman comments that the current draft of the EU AI Act would be over-regulating. While OpenAI would first try to comply with the regulations in Europe in the future, they would consider leaving Europe if they could not comply.
  • Andy Warhol’s estate lost its US Supreme Court copyright fight with the celebrity photographer Lynn Goldsmith, ruling that Warhol’s works based on Goldsmith’s 1981 photo were not immune from her copyright infringement lawsuit. The result of this case may have implications regarding the US legal doctrine called fair use, which promotes freedom of expression by allowing the use of copyright-protected works under certain circumstances without the owner’s permission. As the exceptions and limitations of US fair use have been widely discussed but not clarified in the context of AI, specifically the use of copyright protected works to train AI models, this ruling could have wide reaching implications for AI developers.

June 2023

AI music tech news

  • Octiive announces the development of a suite of tools called ACE that use generative AI to help artists focus on creativity, including one that lets you leapfrog to future versions of a song. Interestingly, Octiive is billed as the largest digital music distributor and creativity hub run by musicians for musicians.
  • Meta launches MusicGen, a music generating tool that can turn a text description into about 12 seconds of audio. MusicGen was trained on 20,000 hours of music, including 10,000 high-quality licensed music tracks, and 390,000 instrument-only tracks from Shutterstock and Pond5 stock media library.
  • ByteDance (TikTok) announce the release of Ripple, a new free-to-use music production app that has two key features: a melody-to-song generator and a virtual recording studio. ByteDance wants content creators to generate music for their short-form videos using the app.
  • WAVs, an AI-music directory, raises a $20m funding round. Music Ally reports that WAVs AI turns out to be “a Spotify/iTunes-style front end” for tracks by artists such as DrAIk, WKND, Ariana Granday, Freddie MercurAI, RhiannAI, Ed ShAIran, etc. WAVs describe the startup as “making compromises with record labels and artists… to provide users with personalised creations of music.
  • Oregon’s Live 95.5 becomes the world’s first radio station with an AI DJ, using cloned human voices to host segments.
  • Microsoft’s AI music project Muzik announces DeepRapper, an AI-powered rap generator.
  • AI company Amadeus Code, known for its songwriting app, releases MusicTGA-HR, one of the largest music data sets in the world and signs a deal with Roland Corporation, a global electronic musical instrument manufacturer and developer, as its first customer. This would allow users of Roland hardware to access Amadeus Code’s dataset for royalty free samples.
  • South Korean digital Service Provider Genie Music announces genie.Re.La, a new service which offers consumers the ability to generate digital scores instantly when they upload an MP3 file. This has been enabled through a partnership with AI music company Juice. Genie Music have commented that only officially distributed music can be uploaded and the created works will be sorted on the company’s server.

Music industry news

  • Grimes partners with TuneCore to distribute AI songs that use her voice through her Elf.Tech Ai voice model. This provides the necessary framework for Grimes to monitor and consent to what songs are released using her AI voice.
  • It is announced that “only human creators are eligible to be submitted for consideration” for a Grammy award. A “work that contains no human authorship is not eligible in any categories”. “The human authoriship component of a work must be meaningful and more than de minimis”. Use of AI as a tool is therefore permitted but any 100% AI-generated work is not eligible. The lines are still blurred though!
  • The Recording Industry Association of America (RIAA) asks Discord to remove AI hub server hosting copyright protected music via a Digital Millennium Copyright Act (DCMA) subpoena. The subpoena requests that Discord disclose the identities of users who had posted Discord links featuring datasets of copyright protected songs.
  • Music licensing agency Rightsify, who specialises in providing curated background music, launches the Global Copyright Exchange (GCX), the world’s largest ready-to-go dataset that spans the globe and is cleared for use. The dataset hosts millions of tracks and robust metadata with more than 400 parameters.
  • Electronic musician and audiovisual artist Damien Roach aka patten releases Mirage FM, the first LP made entirely from AI-generated sound sources through the neural network Riffusion.

Law and regulation

  • Japan’s Personal Information Protection Commission issues a statement asking OpenAI to minimise the sensitive data it collects for machine learning purposes, stressing not to do so without people’s permission. It also warned that it is prepared to take further action if there are additional issues.
  • UK Music calls for the UK government to protect the music industry from AI technology and prevent music laundering, “a process where you could steal someone’s work, feed it into an AI, and then generate clean “new” music”.

July 2023

AI music tech news

  • Epidemic Sound, one of the most popular music libraries used by YouTubers, launches ‘Soundmatch’, a new AI-powered feature that will analyse frames in a creator’s video and then suggest music that is likely to match the content.
  • AI music platform Mubert announces that its AI has generated 100 million tracks. Mubert says the audio files were generated using licensed music for input and that the company has established relationships with music creators who contributed samples to allow Mubert to create a database of 2.5 million proprietary sounds on which to train its algorithms.
  • American musician and DJ BT shares that he has been working on a number of AI music tools that have inspired his new album.
  • Aimi’s app can now do trap music. Trap fans can listen both continuously and interactively to engage with an ever-evolving mix of instrumental trap and hip hop sounds.
  • Google and Osaka University publish research that demonstrates a technique for converting thoughts into music through an AI model called Brain2Music. They hope that this could lead to intelligent music composition software that would be able to decode thougts and translate them into music.
  • Diandre Ruiz of music group Bandlez launches Soundry, a generative AI sample program. Soundry promises to generate on-the-spot samples from prompts with stylistic descriptors, instrumentation and the desired musical key.
  • Meta launches AudioCraft, a free generative AI text-to-audio and music tool. AudioCraft consists of three models: MusicGen, AudioGen and EnCodec. Meta wants to enable musicians to begin exploring new compositions without having to play a single note – generative AI for audio made simple and available to all.

Music industry news

Law and regulation

  • US District Judge Berul Howell in Washing DC reaffirms that only human authors can receive copyrights. An application filed by computer scientist Stephen Thaler on behalf of his DABUS AI system was rejected.
  • Comedian and actress Sarah Silverman files class-action complaints against OpenAI and Meta for training their AI products on copyright protected works without permission. Silverman alleges both direct and vicarious copyright infringement as well as violations of California competition law.

August 2023

AI music tech news

  • Google DeepMind announces the SynthID watermark for AI-generated images. The watermark is impossible to see in an image, but easy for the detection tool to spot. It is likely this will be developed further to detect AI-genreated music.
  • Futureverse joins the realm of AI music with the debut of JEN1, a universal high-fidelity model designed for text-to-music generation. JEN1 is said to outperform Google’s MusicLM and Meta’s MusicGen with superior music quality.
  • Loudly goes to market, giving users the option to create a soundtrack and customise it by exporting it to any DAW of their choice. Users can select instruments, duration, energy level, and genre. There’s even an option for genre-blending.
  • Myvox launches to offer musicians the opportunity to licence human-sounding AI singers for their songs, a legal alternative to voice cloning. “Users can create original songs with these AI-cloned vocals, distribute directly to all streaming platforms, collect royalties and share in the revenue with the artist”. The software debuts with Iranian-Dutch artist Sevdaliza.
  • Beat buying and selling platform BeatStars unveils AI music creation tool, Seeds. The launch comes from a recently announced strategic alliance with AI music startup Lemonaide, which has developed an AI-powered plugin that generates ideas for beats.
  • Moises, an app used to isolate vocals and instruments in tracks, announces three new tools: AI lyric transcription, advanced chords, and song sections.
  • Beat-Forge-AI reveals a music generation application that leverages the power of GPT-3 with AWS, React, TypeScript and Java Spring Boot.
  • Sample editing software SampLab releases a new plugin which lets you generate samples from text for free.

Music industry news

  • Jukedeck founder later acquired by TikTok, and current VP of Audio at Stability AI, Ed Newton-Rex, releases his own original composition featuring lyrics generated by OpenAI’s GPT-3. The work will be published by Boosey & Hawkes.
  • Google and Universal Music Group (UMG) are in talks to licence melodies and voices for songs generated by AI. Discussions are at an early stage but the goal is to develop a tool for fans to legitimately create new tracks and pay the owners of the copyrights for it. YouTube announce the Music AI Incubator, which incorporates inaugural UMG members, the Frank Sinatra Estate and ABBA’s Bjorn Ulvaeus as participants.
  • YouTuber Dopfunk uses AI to create a posthumous collaborative song, named ‘Facing Death’, between late rappers DMX and Tupac Shakur.
  • Electronic music artist Don Diablo releases the first official music video fully created with AI. Set in a dystopian universe, the video features an original track produced by Diablo based on a viral piano masterpiece called ‘Rain’ by Tony Ann. In the video, two protagonists resist an empire controlled by AI, challenging the dominance of technology and highlighting the strength of human spirit.

Law and regulation

  • New York Times considers legal action against OpenAI to protect the intellectual property rights associated with its reporting. The newspaper and OpenAI have been locked in negotiations over reaching a licensing deal in which OpenAI would pay the Times for using its stories, but discussions have broken down.
  • The UK’s Culture, Media and Sport Committee urges government to address threates to creativity from AI, suggesting that government should mediate relationships between AI developers and rights management companies.
  • US Copyright Office launches a study into AI and copyright. The study will examine three topics: those involved in the use of copyrighted works to train AI models; the appropriate level of transparency and disclosure with respect to the use of copyrighted works; and the legal status of AI-generated outputs.

September 2023

AI music tech news

  • Pixelynx launches BeatKOR in partnership with Beatport. BeatKOR will allow creators and fans to remix music from iconic and emerging artists through new interactive modes. BeatKOR is Pixelynx’s latest digital companion that creators and fans can use to create AI-generated music on the KORUS platform.
  • Kits.AI turns a simple hum into a massive orchestra. Built by Arpeggi Labs, Kits.AI lets you turn voices into other voices and sounds into instruments through its Instrument Converter.
  • Spotify launches ‘Daylist’, an AI-powered playlist that updates multiple times per day based on past listening habit for a “hyper-personalised” listening experience.
  • Stability AI releases StableAudio, a tool that Stability claims is the first capable of creating high-quality, 44.1kHz music for commercial use via a technique called latent diffusion. Stability’s audio team relied on previous versions of generative AI music models, creating a new model inspired by DanceDiffusion to underpin StableAudio, which Harmonai then trained.
  • MUSIXY.ai, the world’s first AI music label, unveils the world’s first streaming platform exclusively for new and covered hit songs with unofficial AI-generated vocals of iconic living or deceased singers. John Lennon and Paul McCartney are among the headline acts on the site with the track ‘Peace’ (profits from which will be donated to a Ukraine charity apparently).
  • TikTok launches a new tool to help creators label their AI-generated content. TikTok are also testing ways to label AI-generated content automatically.
  • YouTube launches new AI-assisted tools for creators. Five new features are available, including Dream Screen, which allows people to add AI-generated images and video backgrounds to their YouTube Shorts videos, AI Insights, which suggests video ideas based on what your audience is watching, Aloud, which is an automatic dubbing tool to help creators create content in more languages, an AI-assisted search tool, and a new video-editing and production app called YouTube Create.
  • Suno AI launches on Discord, designed to enable creatives and developers to generate hyper-realistic speech, music and sound effects from scratch. Currently in its beta release stage of development, Suno AI has the potential to power personalised interactive experiences acrpss various platforms, including gaming, social media, entertainment and more.
  • Startup YourArtist.AI offers chatbots based on famous artists. Chatbots promise to sing tracks submitted as a YouTube link or audio file. There’s no fee to chat with a virtual singer, but tehre are fees to sing cover songs or clone your own voice.
  • SoundVerse releases a chatbot assistant as the interface for making music. Users can type messages to the assistant explaining what they want to create.
  • StaffPad, a composition app sold by Muse Group, has released a new feature called Piano Capture. The feature listens to what you’re playing on a piano and converts it into readable sheet music so that composers can capture their musical ideas as they play.

Music industry news

  • Warner Music Group signs a record deal with the first AI virtual pop star. Noonoouri was created in 2018 as an 18-year-old metaverse avatar, before going on to star in fashion campaigns for Dior, Balenciaga and Valentino. She has more than 400,000 followers on Instagram and is set to release her first single ‘Dominoes’, which already has more than 60,000 views on YouTube. Warner have faced some backlash for sexualising a child-like body.
  • Ghostwriter’s fake Drake and The Weeknd track ‘Heart on My Sleeve’ is submitted for Grammy consideration in the categories of best rap song and song of the year. Recording Academy CEO Harvey Mason Jr. expressed his support for the track’s eligibility, arguing that it qualifies due to the human involvement in its creation. Many question whether the song is eligible as it was removed from streaming services due to a copyright violation, and according to Grammy rules songs must have “generation distribution” andhave broad availability through various platforms to be eligible. Harvey Mason Jr later took to social media to clarify that the fact that ‘Heart on My Sleeve’ was eligible for a Grammy was “innacurate information”. This particular version of ‘Heart on My Sleeve’ is not eligible for Grammy consideration as the vocals were not legally obtained.
  • Universal Music Group (UMG) and Deezer announce an artist-centric streaming model, designed to better reward the artists and the music that fans value the most, and to crack down on streaming fraud.
  • Ghostwriter releases another AI-powered song, this time mimicking rappers Travis Scott and 21 Savage. The track is called ‘Whiplash’.
  • Rapper Fivio teams up with Soundraw to generate an AI instrumental for his track ‘Doin Me’.
  • The Canadian Academy of Recording Arts and Science who look after the Juno Awards have announced that recordings that use AI can be eligble for an Juno award, but AI can’t be the “sole or core component of the song/album”.
  • Spotify CEO Daniel Ek states that Spotify will not ban music made by AI. Spotify currently does not allow its content to be used to train a machine or AI model.

Law and regulation

  • Michael Chabon and four other acclaimed authors bring class action lawsuits against Meta and OpenAI, claiming copyrighted works by themselves and other writers were used to train the companies’ large language AI models.
  • Songwriters, incluidng Jimmy Jam, Terry Lewis and Paul Williams are to meet with Congress in a bit to set rules for AI use, advocating for fair limits on AI use as part of ASCAP’s Stand With Songwriters day. The songwriters will urge Congress to adopt legislation on AI in music that adhere’s to ASCAP’s six principles for AI, which include human creators first, consent and compensation.
  • A revised version of US Representative Deborah Ross of North Carolina’s Protect Working Musicians Act has been introduced, endorsed by the American Association of Independent Music (A2IM) and the Artist Rights Alliance (ARA). The new legislation would allow independent artists to band together and collectively negotiate with alrge streaming platforms and AI developers.
  • Writers Guild of America publishes terms of a new Minimum Basic Agreement (MBA): “AI can’t write or rewrite literary material, and AI-generated material will not be considered source material under MBA, meaning that AI-generated material can’t be used to undermine a writer’s credit or separated rights”… “A writer can choose to use AI when performing writing services, if the company consents and provided that the writer follows applicable company policies, but the company can’t require the writer to use AI software (e.g., ChatGPT) when performing writing services”. Songwriters and musicians don’t have the same collective-bargaining structure as screenwriters (in the US at least) but these guidelines could influence deals within the music industry in the future.

October 2023

AI music tech news

Music industry news

  • Alcohol brand Bacardi announces plans to release an AI powered album called ‘The A.I. Powered Album’ with Grammy award winning producer Boi-1da and artists Bellah, Blackway, Floyd Fuji, Kyle Dion and Savannah Ré.
  • Five AI personalities take over Fly FM, Malaysia’s first radio station to exclusively use AI radio DJs for the day.
  • Universal Music Group announces a strategic alliance with generative AI music company BandLab. The alliance promises to advance the companies’ shared commitment to ethical use of AI and the protection of artist and songwriter rights while also promoting market-led solutions with pro-creator standards to ensure new technologies serve the creator community effectively and ethically.
  • Dutch electronic music label Spinnin’ Records teams up with Endel to produce 50 AI-powered sleep and wellness albums to be released weekly.
  • Universal Production Music, a division of Universal Music Publishing Group, has launched a new subscription music licensing service that provides content creators with claim-free music and sound effects for online videos. The new service gives consumers unlimited access to a pre-cleared music and sound effects library, starting at $5.99 per month. With a Creator Pro tier priced at $12.99 per month for commercial subscribers, this move places Universal Music firmly in the market place to compete with AI production music in the future.
  • As part of Warner Music Group’s deal with Endel, Roberta Flack’s ‘Killing Me Softly’ gets reworked into three soundscape albums to suit different activities: focus and productivity, relaxation and sleep.

Law and regulation

  • US digital rights organisation Fight for the Future partners with music industry labour group United Musicians and Allied Workers to launch #AIDayofAction. This campaign calls on Congress to block corporations from obtaining copyrights on music and other art made with AI.= in order to prevent major labels and other music industry businesses from being able to copyright protect works generated by AI, thereby safeguarding human creativity.
  • The Independent Music Publishers International Forum (IMPF) proposes four ethical guidelines for AI developers: seek permission for the use of music; keep records of the music used to train AI; label AI-generated works; and delineate between AI-assisted works and 100% AI generated works.
  • UK government-backed Innovate UK launches £1m funding contest for the development of AI music, to advance the development of Artificial Intelligence products and services within the global music supply chain which benefit and strengthen the UK Music Sector.
  • The Recording Industry Association of American (RIAA) adds voice cloning to its Notorious Markets Report as a threat to artists and copyright owners.
  • US senators Marsha Blackburn, Chris Coons, Thom Tillis and Amy Klobuchar launch the No Fakes Act (Nurture Originals, Foster Art and Keep Entertainment Safe Act). The proposed bill would “protect the voice and visual likenesses of individuals from unfair use through generative artificial intelligence”.
  • Music publishers Universal Music, ABKCO and Concord Publishing sue Amazon-backed AI company Anthropic over the use of copyright protected song lyrics to train their chatbot Claude. This marks the first lawsuit led by the music industry.
  • US President Joe Biden publishes an executive order on “the safe, secure, and trustworthy development and use of artificial intelligence”. A spokesperson for the Human Artistry Campaign commented: “The inclusion of copyright and intellectual property protection in the AI Executive Order reflects the importance of the creative community and IP-powered industries to America’s economic and cultural leadership,”
  • UK Music interim chief Tom Kiehl writes to UK Primeminister Rishi Sunak to urge him to consider the implications of AI on the creative industries at November’s AI Summit to be held at Bletchley Park.

November 2023

AI music tech news

Music industry news

  • The world premier of Denis Woychuk’s AI musical, ‘The Duchess of LES: An AI Musical From Little Ukraine’ is held at the Kraine Theatre in the East Village of New York City. This production blends AI music created by programmer/composer Kyle Williams and original music created by composer Darrell Lawrence.
  • The Beatles release their final song ‘Now and Then’, which uses AI to remaster John Lennon’s voice from demos recorded by Lennon and George Harrision’s guitar playing. Paul McCartney, George Harrison and Ringo Starr working on Now And Then in 1995. Film director Peter Jackson was instrumental in developing the audio restoration technology that allowed the extraction of Lennon’s vocals.
  • Believe, Paris-based parent company of TuneCore, a music distribution service, confirms it has developed and launched a proprietary “AI detection algorithm”, via a tool called AI Radar. Believe claims the tool can detect AI-made music with 98% accuracy.
  • The estate of Edith Piaf partners with Warner Music Entertainment and the production company Seriously Happy for an upcoming biopic that will use AI technology to recreate the late singer’s voice.
  • Recently, The Orb and David Gilmour, the guitar and voice of Pink Floyd released Metallic Spheres In Colour, described by Sony Music as “a visionary re-interpretation” of their 2010 ambient album Metallic Spheres. Now, Sony Music and the artists are inviting fans to use AI to remix the music from the new album and its cover art, using AI-powered tools to create their own personalised music track and artwork.

Law and regulation

  • ASCAP comes out strongly against AI training as fair use in its submission to the UC Copyright Office inquiry into the copyright issues raised by generative AI technologies. “The AI industry can and should obtain consent and pay fair market value for copyrighted musical works before using them in the development of AI models”. ASCAP also spoke against the idea of compulsory licensing due to its inefficiencies and shortcomings.
  • The UK hosts the AI Safety Summit at Bletchley Park. UK Music interim chief Tom Kiehl urges UK prime minister Rishi Sunak to use the summit to protect music creators and sector’s “fragile” talent pipeline.
  • At the OpenAI ‘Dev Day’ conference, the CEO Sam Altman announced that the company will act as a “Copyright Shield”: they will step in and defend their customers and pay the costs incurred, if they face legal claims around copyright infringement. This applies to developers using the ChatGPT Enterprise product and OpenAI’s API.
  • YouTube announces that it will soon require creators to label their content as AI generated, and will tell users when they are watching content generated by AI.
  • Following the filing of a copyright infringement lawsuit back in October, Universal Music, Concord Music Group and ABKCO Music ask the court to require Anthropic to “implement effective guardrails” that would prevent the company’s AI models from reproducing or distributing the copyrighted song lyrics, and to prevent the company from using these works to train future AI models.

December 2023

AI music tech news

  • AI music startup DAACI launches ‘Natural Series’, a comprehensive product portfolio of plugins, music tools and editing technology. The series kicks off with ‘Natural Edits, a music editing tool for the global sync market to make any track instantly adaptive.
  • AI music company DigiTrax unveils their new AI Music Training Model License offered in its patented AI platform KR38R LAB’s Artist and Label Services. Base on seven patented AI-powered music composition technologies and solid music theory foundations, KR38R LAB utilises abstract music theory data extracted from composition analysis to generate music theory data block templates that encapsulate a composer’s unique style. Once an agreement is signed between DigiTrax and artists, the ‘Artist Authorized’ AI Training Template is generated by KR38R LAB under this license. These templates serve as foundational elements for music creation, allowing new music creators to remix and utilize them in KR38R PRO or other production tools, fostering diverse sonic landscapes.
  • Google Arts & Culture Lab team, launches ‘Instrument Playground’ uings the MusicLM model that Google launched earlier this year. Users can choose to hear 20 second clips of over 100 instruments from around the world in any mood or style they wish.
  • Girls Who Code, a nonprofit working to close the gender gap in tech, launch GirlJam, an AI songwriting experience empowering girls to get comfortable with and explore the creative possibilities AI technology can offer.
  • Spotify downsizes by laying off 17% of its workforce, while focusing on delivering a hyper-personalized service and investing increasingly in artificial intelligence.
  • Music licensing firm Rightsify has announces the opening of its datasets and AI models to allow anyone to generate high-quality AI music from just text prompts. The dataset of 60,000 songs with more than 300 unique musical instruments has been trained exclusively on the Rightsify-owned catalogue, making the company the first music rights holder to turn its expansive music library into an AI model open for everyone to use.
  • Meta launches AudioBox, an innovative tool which allows users to convert text into speech, compose music, and create sound effects with using simple text prompts.
  • Google’s AI Test Kitchen launches a new AI music generator, MusicFX. Users can write prompts to generate music lasting from 30 to 70 seconds and can loop back to the beginning. Users can download and share the result, however the end product will includes a unique digital watermark from from SynthID developed by Google subsidiary DeepMind.
  • AI-powered music startup Mubert introduces a new image-to-music feature.
  • Microsoft partners with AI music company Suno to bring AI-based music generation capabilities to AI-powered chatbot Copilot.

Music industry news

  • Endel partners with Sia to work on her EP ‘Gimme Christmas’, putting the festive tracks through its AI to create two ‘soundcapes’ to be released on streaming services as well as within its own app.
  • American rock band KISS Rock band KISS debut a “new era” for the group, unveiling their digital avatars during the “End of the Road” farewell tour. The avatars performed their iconic “God Gave Rock and Roll to You” during their encore.
  • Hip-hop artist Nimrod Cain, or “Nimsins”, records the first studio album produced by generative AI using MusicGen technology developed by Meta. The album recording process, consisting of 8 songs recorded over a 24-hour period, has been catalogued in a two-part documentary ‘Sensory Overload: The AI Music Producer’ available via Slouchy Media.
  • Warner Music Group has partnered with British health tech startup MediMusic to explore the potential of music as a therapeutic tool. Their algorithms analyse music’s ‘digital DNA’ to create a healthcare-specific fingerprint, to curate 20-minute personalized tracks, delivered through the MediBeat streaming device and a pair of headphones. The running order of playlists curated by MediBeat is designed to reduce heart rate and stress hormones, such as cortisol, while promoting relaxation through the release of hormones like dopamine and oxytocin. MediMusic monitors physiological responses in real-time via a wrist heart rate monitor.

Law and regulation

  • The three main institutions of the European Union – its Council, Commission and Parliament – near the end of their negotiations to agree a final draft of the EU’s new AI Act. Leaked text of the proposal suggests the Act will require foundational model makers to put in place a policy to respect EU copyright law, including with regard to limitations copyright holders have placed on text and data mining, and to provide a sufficiently detailed summary of training data used to build a model and make it public.
  • In a new submission to the United States Copyright Office, ASCAP doubles down on support for direct voluntary licensing and a Right of Publicity expansion.
  • The New York Times files a copyright infringement lawsuit against OpenAI. The Times’ action also names Microsoft as a defendant. In April of 2023, the New York Times “reached out to Microsoft and OpenAI…to raise intellectual property concerns and explore the possibility of an amicable resolution, with commercial terms and technological guardrails that would allow a mutually beneficial value exchange,”. As no resolution was reached, a lawsuit was the newspaper’s last resort.
  • In the case of Thaler v Comptroller General of Patents Trade Marks and Designs, the UK Supreme Court rules that AI cannot be an inventor for UK patents. Within the meaning of the Patents Act 1977 (the “1977 Act”), an inventor must be a natural person who devises a new and non-obvious product or process (the invention). Any inventions entirely conceived by Stephen Thaler’s AI DABUS cannot receive a patent in the UK, because a person cannot be identified as the inventor.

Leave a Comment