Jeremiah Krage

Will AI replace musicians? The music industry has undergone significant technological advancements over the years, from the introduction of analogue gears to the widespread adoption of Digital Audio Workstations (DAWs). However, one of the most groundbreaking developments in recent times is the integration of artificial intelligence (AI) into music creation processes. We all have heard the fake viral songs featuring AI versions of Drake, The Weeknd, & Sidhu Moosewala, which were created using different AI platforms, and one cannot deny the fact that we all enjoyed those fake versions. Also, there are platforms that are already creating AI-based music, and music producers are already using it. This raises the question of concern: Will AI replace musicians?   Reflecting on the past, the introduction of the first MPC machine in 1988 stirred concerns among session musicians about being replaced. However, history has demonstrated that technology serves as a tool to enhance rather than replace human creativity. MPCs not only saved time for music producers but also inspired many session musicians to delve into music production themselves. Moreover, AI is making significant contributions to the music industry. One notable example is how AI facilitated The Beatles in completing a new song after 45 years. Utilizing an old demo tape by John Lennon, AI algorithms helped reconstruct and finalize the track, showcasing the potential of AI to resurrect musical works.   Pros: 1. AI can swiftly generate multiple musical ideas, offering valuable inspiration for music projects 2. AI streamlines the music creation process, enabling music producers to meet deadlines more efficiently and cost-effectively 3. Royalty-free AI voices provide a platform for exploring different vocal styles, facilitating experimentation and creativity Cons: 1. There is a risk that reliance on AI-generated content may discourage musicians from developing their own melodies, potentially leading to a decline in human creativity 2. AI-generated music may lack the emotional depth and authenticity that human musicians bring to their compositions, potentially resulting in a sterile or mechanical sound 3. The unethical use of AI for creating deep fakes and replicating famous singers' voices raises ethical dilemmas within the industry, necessitating careful consideration of the ethical implications of AI-generated content Final Thoughts: While AI holds immense potential to revolutionize the music industry, it is unlikely to fully replace human musicians. AI should be viewed as a complementary tool, augmenting human creativity and offering new avenues for exploration and innovation. AI-generated music should be ethically sourced and properly licensed, ensuring fair compensation for creators. By striking a balance between human artistry and technological innovation, we can continue to push the boundaries of musical expression and create transformative experiences for audiences worldwide.

3

2 Steps Forward, Half Step Back… Creating a successful piece of IP is hard work no matter what platform you target. YouTube is no exception. And while YouTube has its own special set of challenges (namely, an opaque algorithm), it also gives creators the unique ability to experiment in real-time with their audience. I’ve talked before about how YouTube provides reams of analytics on every video to help you figure out what’s working and what’s not. I love looking at the data and find it helpful as both a serotonin booster (wow, look how many people like our video! We better keep making them!) and to understand where we need to improve. But I’m always wary of misinterpreting noise for signal. That’s a slightly techy way (hey, I’ll always be an engineer at heart!) of saying that it’s easy to read too far into the data and draw unwarranted conclusions. For example, we’ve released a couple of videos that didn’t perform well over the first 24 hours, so we’ve taken them down, tweaked something small, and re-released them. And until this weekend, we’ve always had much more success on the second upload. Often by a factor of 10 or more. But this weekend, we had our first flop. We released our 8th video early Saturday morning, and when we saw no uptake all day Saturday, we took it down, made some fairly significant changes to the video (plus thumbnails, titles, descriptions and keywords), and tried uploading again Saturday night. Our performance was even worse! It was frankly a bummer. We’ve seen nothing but straight-line improvement on all metrics for months, but it’s also a good lesson. Past success does not guarantee future progress. But because we released the video on YouTube, we see this failure as a chance to experiment. We took down the video again and at some point soon we will re-release testing out new theories about why the algorithm didn’t serve it up to more people. We have a strong inkling the video didn’t resonate because we started this video with strong negative (sad) emotions before presenting the fun and positive solution to our kid’s problem, but we can’t be sure. Perhaps when we post the third time, it will perform like our other videos and we’ll draw a different conclusion. Or more likely, we'll steer clear of introducing anything negative in our song intros in the future! Either way, we'll have learned something. In the meantime, we continue production on more videos and celebrate the growth we have achieved this week. We are now over 900 subscribers and over 250,000 total views. Not bad for our fledgling, no-budget channel. Have questions? Drop me a comment. I’m an open book. #YouTube #indieanimation #brandbuilding

22

2 Comments

Exciting new research from Microsoft explores VASA-1, a system capable of generating incredibly realistic facial expressions and head movements synchronized with speech audio. This extends beyond unseen talking faces in photos to even handle singing voices, demonstrating its versatility. Key advancements include: ❇️High-fidelity lip-syncing with natural facial dynamics. ❇️Real-time generation of high-resolution (512x512) videos at 40 FPS online. ❇️Minimal startup delay for smooth virtual avatar interactions. The impact? VASA-1 has the potential to revolutionize real-time interactions by enabling lifelike virtual avatars that can engage in natural conversation. This technology has the potential to create more engaging virtual assistants, but there are concerns that it could be misused to create deepfakes. What are your thoughts on VASA-1 and its potential applications? Share your comments below! 👉 Source: https://lnkd.in/gMdRCBhB 👉 Follow me if you're also exploring how Generative AI is changing the way we create, design and tell stories in art, design, entertainment, and beyond. https://lnkd.in/dtRMyMHW #generativeAI #AI #VAs #deepfakes #MicrosoftResearch

2

There's much hyperbole and a saturation of immature products in the current music AI space. A few years back I was pitched daily by Web3 / NFT startups (remember those?) promising to reinvent the music industry and leave sceptics trailing. Their technology was mostly brilliant, but several years on and the revolution we were sold isn’t close to fruition. We've reached a similar point with gen music AI. The same folks who vied their Web3 startup would take down major labels are now telling us 'traditional' music creation will die a painful death. If Suno et al are creating a future where the music creator and listener become one, this gives serious potential to disrupt DSPs, aggregators and everything in between. Welcome thinking, but there's a long road before such utopia can be meaningfully construed. Not disclosing training sources will likely hurt these services. Brand trust begins with responsible practice. It's feasible high profile artists will boycott and publicly vilify them - think SAG/AFTRA strikes but with music. Suno’s millions will aid with legal challenges, but could set an uneasy precedent where others follow - steal now, seek permission later. Will future legislation render AI music services a gimmick unfit for public broadcast or distribution? End user agreements for most gen AI services are unacceptably vague. We see “Use at your own risk” through to “your creation is uniquely your copyright”. An hour spent researching popular service EULAs gives me little comfort for pro/broadcast use or publishing to a DSP. Output quality is key: An audibly pleasing, musically interesting output is de rigueur. Weaving drum hits together with a browser based MIDI generator is not a product that will disrupt the face of music creation. Buying a bona fide (human created) output may be far more valuable than hours spent prompting a turd. With a glut of products to choose, UX and integration into existing creator workflow are massive: If UI is poor, it takes too long to generate an output or input fields yield inaccurate results, interest is rapidly lost. Is there an exaggeration of AI’s impact? We will doubtless evolve an entirely new class of creator, further democratising music making. After all, why settle for spoon-fed major label playlists when you can just roll your own bespoke experience? But the threat to professional creator livelihoods seems a long way off. The winners in this space will likely be the well-researched, steady adopters who subtly integrate AI assisted features to aid existing creative processes and democratise the learning curve for tomorrow's creators today.

108

18 Comments

The first episode is out there. (with English subtitles) Watch it & if you like it, I need a shout out. #wirally #mansionhouseboyzzzz #comingsoon #newseries #mansionhousebrandy

3

To many, it was just a throwaway diss. To me, Metro Boomin's 'BBL Drizzy' was one of the most important milestones in the story of AI and Music. Why? BBL Drizzy is an instrumental beat by Metro Boomin, released during the so-called civil war of Hip Hop to point attention to Drake's *ahem* alleged surgical enhancements. It sampled a song that... never existed. Over a mid-70s sounding Soul/R&B arrangement, a singer from a bygone era proclaims 'the best BBL in history'. It's not revolutionary in and of itself, but I believe BBL Drizzy demonstrates a sea change in music: 1. AI is now making plausible sounding automated samples and can replace the sampling process of old: - Sampling is a huge industry in music, particularly hip-hop. It can be expensive and one mistake can have material ramifications; Automating samples clears this hurdle, making sampling cheap, quick and efficient - On the flip side, sampling breathes life into old records, providing a catalogue artist a payout and the record a second wind. Automated sampling *will* take money away from catalogue artists 2. AI makes tasks that are complex to humans really, really simple - Retrofitting samples has been done before (see Tuxedo's 'Number One') but it takes significant songwriting hours, a band, a recording studio and mixing/engineering time - AI takes a few prompts and runs with it - Likewise, the Beatles were able to finish a final song ('Now and Then') they previously couldn't by using AI to isolate vocals from the backing track and 3. BBL Drizzy shows AI democratise music - BBL Drizzy's sample was made by a comedian (King Willonius), not a professional musician; AI opens up music production to *anyone* with an idea; widening the possible pool of artists - Owing to the lack of copyright restrictions, Metro Boomin was able to crowd-source verses to the track, allowing literally any rapper to rap on the BBL Drizzy beat 4. AI can devalue music and disempower artists - AI can cut the skilled professional out of the equation; by automating drums, guitars, bass, etc. AI risks removing the experienced, trained professional from the equation, and the money pool - AI tools can correct or improve a performance, possibly reducing the need for practice and refinement - this might lead to a general reduction in the average skillset, hurting artists on the live circuit We're likely to see unprecedented investment into music-based AI over the coming years and it's my view that we need to pre-empt the AI-sized train coming through the tunnel with regulation. UK Music has set out 5 principles for AI in music (https://lnkd.in/esDfpNU3), but it's my view that we need a Rome Convention-type response to extract and leverage the benefits of AI and to allow a new industry to flourish, while protecting artists' income, livelihood and moral/legal rights.

16

6 Comments

Cameron and Kamal Hassan: Pioneers of AI in Cinema The recent news that James Cameron has joined Stability AI's board and Kamal Haasan is attending an AI course is incredibly exciting for those of us who are already experimenting with generative AI in filmmaking. These two iconic figures, known for their groundbreaking work in the film industry, are now embracing the power of AI to push the boundaries of storytelling and visual effects. The Potential of AI in Filmmaking AI has the potential to revolutionize the filmmaking process in countless ways. From generating new ideas and scripts to creating stunning visual effects, AI can enhance the creative process and open up new possibilities for storytelling. As a filmmaker who is already using generative AI, I am particularly excited about the potential of AI to: Create unique and original content: AI can generate new ideas and storylines that would be difficult or impossible for humans to conceive. Enhance visual effects: AI can be used to create stunning visual effects that would be prohibitively expensive or time-consuming using traditional methods. Improve efficiency: AI can automate many of the repetitive tasks involved in filmmaking, freeing up filmmakers to focus on more creative aspects of their work. The Role of Industry Leaders The involvement of industry leaders like James Cameron and Kamal Haasan in the AI space is a significant development. Their influence and expertise can help to legitimize the use of AI in filmmaking and inspire other filmmakers to explore its potential. Looking Ahead As AI technology continues to advance, we can expect to see even more innovative and groundbreaking films being made. The collaboration between filmmakers and AI experts will be crucial in shaping the future of cinema. I am excited to be part of this exciting new era and to see what the future holds for AI in filmmaking. ISR Selvakumar #StabilityAi #Kamalhassan #JamesCameron #Ai #AIMovie

12

Yanina Strylets & Viktoriya Strylets of SciMatch unveil their AI-driven app, tackling dating fatigue & fostering deeper connections beyond looks. Full interview at kingofpodcasts.com or watch on my YouTube and please like share and subscribe! #DatingApp #AI #SCiMatch

10

1 Comment

In a groundbreaking move, Voice-Swap, an innovative AI voice company co-founded by DJ Fresh, has launched a free-to-use VST plugin that's set to revolutionize music production. This cutting-edge technology allows producers to seamlessly integrate licensed voices into their digital audio workstation (DAW), with automatic royalty payments to the artists behind each voice. With Voice-Swap, producers can choose from a menu of iconic voices, including house music legend Robert Owens, Ayah Marar (Calvin Harris and Nervo collaborator), Angie Brown (The Rolling Stones, Happy Mondays, Grace Jones), and more. The plugin enables users to tailor lyrics to fit their creative vision, ensuring a personalized touch. What sets Voice-Swap apart is its commitment to ethical AI use. The company has earned a Fairly Trained Certification, ensuring all current models meet the highest standards. Additionally, custom voice uploads are cross-checked against BMAT's vast database of over 180 million audio fingerprints to prevent unauthorized use. Voice-Swap's CEO and co-founder, Dan Stein (aka DJ Fresh), envisions a future where producers can effortlessly collaborate with their favorite vocalists. "Being able to load up your favorite vocalist straight into your DAW has always been the vision for Voice-Swap," he explains. "Now you can also make your own high-quality models and load them straight into your session." The plugin's custom model training feature allows users to create and share their own voice models, which can be shared privately with other producers, musicians, or labels and agents. This innovative approach democratizes access to high-quality vocal talent, empowering creators to push the boundaries of music production. By prioritizing ethical AI use and fair royalty payments, Voice-Swap is poised to transform the music industry. As Stein notes, "Our BMAT integration ensures rightsholders can trust your models. Once that trust has been established, those models can be shared with other Voice-Swap users, or even made available to licensees and partners via our new API." How do you think AI can enhance music production while promoting inclusivity and fairness? #EthicalAI #MusicProduction #Innovation #MusicTechnology #VSTPlugin #GameChanger #MusicIndustry #FutureOfMusic #AI #Tech #VoiceSwap

10

BREAKING: Universal partners with AI startup SoundLabs Imagine transforming your voice into an instrument. Or another vocal style. SoundLabs' MicDrop can do it. Here's why this AI tech is a game-changer: 1. Voice-to-Voice transformations: Artists can explore new vocal personas. With AI, you can experiment with different styles. And push creative boundaries. 2. Voice-to-instrument conversions: Turn your vocals into instruments. Expand your music production possibilities. Synthesize your voice into guitars, drums & more. 3. Speech-to-singing capabilities: Transform spoken words into sung vocals. Create intricate melodies. And harmonies from your speech or rap. 4. Language transposition: Sing in multiple languages without losing vocal quality. Reach global audiences. No need for multilingual recordings. MicDrop offers ultra-high fidelity vocal models. Ensuring full ownership and artistic control. Artists retain their unique sound. While exploring new creative avenues. ♻️ Repost this to help others in your network. P.S. Can you imagine Jon Bon Jovi use AI? P.P.S. I love his distinct raspy vocals. He's like no other. #ai #artificialintelligence #music #musicproduction #musicindustry ______ If you liked this, join The AI Musicpreneur - my newsletter that helps major artists, creators & industry leaders use AI responsibly to create & promote music like never before. https://lnkd.in/ewhM9U3h

19

12 Comments

In a surprise to no one, Suno & Udio are being sued by the Recording Industry Association of America® (RIAA) for copyright infringement. The RIAA is seeking $150K for every violation. With millions of songs in the catalogs of the majors, you do the math. What's unusual is the consortium who is on board with the lawsuit: the three majors, independent label trade group A2IM, National Music Publishers'​ Association, SoundExchange, SAG-AFTRA and others. The music industry doesn't usually get a ton done, policy-wise, because there are so many trade groups representing so many various factions. The individual groups are usually at odds in any particular dispute. That's not true here; it's rare to have this consensus. What happens now? Unlike years past (see: file sharing), the industry doesn't want these Gen AI platforms to go away. There's clearly consumer demand. The rights holders just want to be able to control/ influence/ monetize them. But this isn't a straightforward case. It's not at all clear that Suno & Udio's use of music as training data will be considered copyright violation. There are a whole range of outcomes- from complete vindication to the services being shutdown. My guess? They settle in a complicated deal. Even though Suno has reportedly raised $150M, that's not enough to pay for content in a traditional streaming/ social model. And with streaming growth slowing, the rights holders are motivated to find the next growth area. This may be it. #musicindustry #musicbusines #musicAI #musictech

50

14 Comments

This is a remarkable film created with AI, and you know why? Because it showcases the art of storytelling—a significant challenge for anyone working with AI films. If you have 10 minutes, sit down and watch it. Here, you can clearly see an alternative way to create fiction, commercials, music videos and TV with AI. The evolution of visual storytelling is happening every day, influenced by every piece of work we do with AI. Sit down and enjoy. well done Digital da Vincis #CreativityAI #AIfilms #CommercialsAI #films #commercials #fictionfilms #FictionAI #ErnestHemingway

6

4 Comments

🎶 You know... there's a way to solve these huge music copyright issues that have been looming over Suno and Udio. 💡 Use the same model that Spotify is using to remunerate creators. 🎧 Every generative AI model is using training data from hundreds of thousands/millions of sound recordings, but no one knows who/what/when/where. 🤷‍♂️ Step 1: AI platforms can be integrated with blockchain-based smart contracts that automatically track and record the use of sound recordings in training data. Every time an AI model utilizes a sound recording, the smart contract mints an ERC-20 token based on minutes, number of iterations, or other metrics and assigns tokens to the associated rightsholders' wallets. 💻🎵 Step 2: Platforms can allocate a portion of their overall proceeds directly back to rightsholders as a percentage of revenue, exactly like streaming platforms do, based on holdings of that ERC-20 token. It's fungible, tradeable, sellable, mintable, and represents exactly the metric that the AI platforms or rightsholders want it to represent. 💰 This ensures not only direct compensation for usage but also continuous revenue sharing as the AI platform grows. 📈🪙 Step 3: Profit? 🎉 Of course, this assumes that platforms want their AI to be responsibly trained, but I'm not going to go down that rabbit hole. 🕳️🐇 #musicai #musictech #musiccopyright

19

20 Comments

While a useful summarizer, using AI to write software is ultimately another pointless attempt at wish fulfillment. Be careful what you wish for alludes to the idea that no prompt can ever be detailed enough to capture the desire properly. That what you get, on the whole, especially as the complexity increases, is increasingly entropic having consumed what seed value its been offered. Only a true and authentic production process actually creating the thing you want has any chance at all of being realized in this world. Ergo, real programmers will always be needed to write original software of any real worth because they are authentic and AI is just a digital simulation of a Jinn in a magic lamp.

3

Has anyone heard about Niantic Studio? Just dropped last week - It's a web-based XR and AR creation tool, and it's really caught our attention here at Tungsten Media. 8th Wall are not new to the AR scene, they’ve been making some impressive AR stuff over the years, but this week they've just released their studio tools for creators to use directly and for FREE. This is going to be a major step forward in making web based augmented reality super accessible and user-friendly. We’ve got to remember - AR is not a new concept but in the last couple of years that barrier to entry has massively shifted to make AR experiential elements in wider campaigns far more easy to integrate. The Niantic Studio system has built in features like: world physics as standard, real-world interactions based on GPS, and even facial filters! This is like having Unity and Effects House, without all the bloat! Imagine this: While we’re working on a giant 3D TV in the centre of Berlin, we could add a REAL (well AR) version of the creative into the physical space on the ground, and capture the passive and active engagement of both worlds colliding. This type of wrap-around creative ideation is the real next level stuff we love to do here! Digital + Physical interactions! We love making innovative processes accessible for everyone, and we’re always exploring the newest tools in AI, CGI, and AR/VR so we are really excited to try this one out. If you’ve got a ‘crazy daft’ idea that fits the bill - we’re the guys to tackle it!

6

2 Comments