The Tritone Paradox hits again, this time for AI

Just a few months back over 1000 musicians released a Silent Album in UK featuring recordings of – guess what- sheer silence. They wanted to protest against the incidence of AI training on musical works without a license. The album, which had 12 tracks, was interestingly labelled ‘Is this what you want’? The question is - even if the answer is ‘No’, what tools and solutions do we have for those who want to be on the right side of fairness and copyright laws? Turns out Meghna Mittal - Co-Founder and CRO at Hoopr, India’s first music licensing platform (as it claims), has some answers. Let’s plug in our earphones to this new track for more..

What made you compose something like Hoopr? Is it pro-AI or fighting against it?

We are a music licensing platform that has introduced India’s first AI-powered features to help brands find the right music fit and generate fresh content ideas for their reels / campaigns. It is built to bridge the gap between content creators/brands and music IP owners, addressing the prevalent lack of understanding regarding music rights in India.

We are like the Quick Commerce for Music. Even ‘using music’ is a market use-case.\

Ours is a first-of-its-kind, self-serve, automated music licensing marketplace- and our idea is to make Bollywood and regional music easily accessible and legally licensable for brands, agencies, and content creators. Our use-case is different from the other use-cases that have prevailed so far. Because even ‘using music’ is a use-case. That’s where the lines get blurred and most of us, including industry professionals, are unable to differentiate between various areas- free, master, publishing and other underlying rights (unlike in the West where structures are clearer).

The source of any AI tool should remunerate the IP owner. A musician’s IP is created through a rigorous process of learning, rehearsing, composing, and producing - making it a valuable asset.

There are influencers and brands that can find a solution here in a structured way – if they want to use someone’s else’s IP for their work. We help them get the licenses and bridge the gap between creators and users. We are like the Quick Commerce for Music. This is a need of the hour as in the minds of many users – music is still something that is free.

‘Music is perceived as free’. Interestingly said and reminded of. Many industry watchers are comparing AI in music to the Napster phase. Has AI become a threat for fair use, creative safety of artists and their monetisation opportunities? Specially with the rise of AI-training on music samples and the inevitable surge of deep-fakes?

With every new technology, guidelines and guard-rails have to follow. The question is – what are you creating it for. Some are doing it for entertainment purposes and creative experimentation. But I don’t think big brands want to risk reputation and legal trouble. We are trying to encourage right and easy monetisation opportunities for artists. And we are also trying to help them accelerate and expand discoverability. Hoopr plans to train AI on its own 100 per cent owned music library if we ever venture into generative music.

Do you make space for both attribution and compensation for the artist?

Yes. And master and underlying rights function for different cases.

Suno (An AI-music generator) has recently come up with a new sample argument (that says that just sounding identical or similar to the original is not enough of an allegation unless a sample from the original exists)- stressing that AI-generated outcomes are different from original works. How do music artists fight such contentions?

We have taken a stand that we will always help IP owners monetise their work better. We use AI to enable artists and with AI-tagging they can connect to new revenue opportunities in an effective and swift way. We are currently not generating any AI music. But the technology is evolving. Today, so many of us use ChatGPT. However, I strongly opine that the source of any AI tool should remunerate the IP owner. I don’t know a solution to many current conflicts and issues around AI and music. The problem with AI-generated music is that no one knows the source and how to remunerate the source. An artist creates music after spending a lot of time, resources, hard effort and creativity. People confuse creating music with using it. Our platform is trying to help both supply and demand side of this industry. Exceptions are there but brands do not usually want to be on the wrong side of law.

There was also the OpenAI Sora’s argument of ‘opt-out’ default. Your take on that?

Well, did the IP owner give the AI tool the files? Did they take any permissions? Then, it’s -by default- not fit for use. We are also concerned with such platforms. For us- the files are given by the music artist.

Can music be treated as a commercial asset?

Music should not be an afterthought for brands and creators, just like any other purchased commodity or service used in content creation. The source of any AI tool should remunerate the IP owner. A musician’s IP is created through a rigorous process of learning, rehearsing, composing, and producing, making it a valuable asset. Platforms like Meta intend their music libraries for personal use only, and commercial usage by brands or influencers requires proper licensing.

What has been the impact of the “20-second rule” for music usage and does the whole space change now?

It is no longer relevant due to advancements in fingerprinting mechanisms. Even 5 seconds of music usage can, now, be recognised, allowing rights holders to detect where their music has been used and for how long. Hoopr has also developed technology to help brands and labels track their music usage, aiming to educate the market about the need for licensing.

In India, ownership agreements are often informal handshake agreements, presenting an opportunity for improvement.

What role can finger-printing play here- specially with ideas like Blockchain, neuralprinting, SoundPatrol tool (Sony and Stanford’s lab)?

With Blockchain, the industry still needs to figure out how to solve the issue of how each piece is tracked and how artists get compensated. It’s a grey space so far. A lot of international tools and platforms are already solving the fingerprinting part for detection of usage. We are also using a combination of some tools.

Do both derivative detection and AI detection get done easily now?

In some cases, derivatives are captured well. In some cases fingerprints are not advanced enough. But in most cases and tools detection happens to the tune of 85 to 90 per cent.

Do watermarks work for music too?

Yes, audio watermarking technology exists, including imperceptible ones, which many labels utilise to protect their content.

How weak or strong is the regulatory plectrum here?

The government needs to talk about this space and adapt for new forces. Many small artists are dying because of lack of shows and gaps in tracking who used their work. IPRS is doing a good job in the area of folk artists. A lot has to come in the form of adequate and strong legal protections. In India, many streaming platforms shut down because of lack of monetisation routes and ecosystems. We can observe the decline in payouts for artists in the digital era compared to traditional music sales, despite increased streaming volumes. China’s mandatory audio streaming subscriptions is an example of how a business model can thrive when users pay for content, specially when we start contrasting it with India’s low subscription rates.

Why is India ready for Hoopr? How does the platform work?

If you look around, presently over 32,000 branded content videos are created daily in India — and that’s across brand handles and affiliate influencers. The production costs are low, but the pressure to stay relevant and culturally aligned is high. Brands are struggling with two issues- one- choosing the right song that aligns with their identity, and second- overcoming creative fatigue when producing content at scale. Hoopr’s AI scans a brand’s digital footprint (Instagram handle, past campaigns, tone, personality, and music choices) and calculates a brand match score with any song on the platform. It helps marketers pick tracks that are not just trending but also authentically aligned with the brand’s identity. Once a brand selects a track, the AI tool generates campaign-ready content ideas that are bespoke for reels, short-form videos, and festive or promotional campaigns.

How do you ensure safety against plagiarism? And do lyricists and other contributors get enough credit and fair compensation too?

We have a strong QC layer and before onboarding we ascertain well about the work’s originality and credibility – automated detection tools also help here. Hoopr currently works with artists and labels known not to have plagiarism issues, and plans to implement automated detection layers before opening the platform to all musicians. There is Metadata associated with each song, provided by the owner, details composers, lyricists, and producers, and Hoopr ensures agreements are in place for song ownership, especially with labels. However, for independent artists in India, ownership agreements are often informal handshake agreements, presenting an opportunity for improvement.

We have to make digital creators and influencers aware that – with technology they can easily stay on the right side of licenses and be fair to artists.

What’s your biggest struggle and achievement – so far?

Creating awareness about this issue. India is still evolving here and we have to make digital creators and influencers aware that – with technology they can easily stay on the right side of licenses and be fair to artists. The general awareness of how the system works for using music in content is lacking, especially for content creators and brands. while some brands are conscious of IP rights, many others adopt a ‘use now and fight later’ attitude. Music is a universal language that helps brands stay relevant across demographics, but a lack of education about music IP rights in India often prevents proper licensing. Traditional TVC licensing, with its high costs, doesn’t suit the daily needs of modern digital content creation, leading to the rise of ‘micro sync’ licensing for organic reels and social media content. So far, we have helped some artists jump the gap to the right opportunities in the area of digital assets. We are the only ones doing what we do- in India. I can think of an example of an artist earning 21,000 rupees from a single sync license, which would require five lakh streams on a streaming platform, demonstrating how Hoopr helps artists who, otherwise, struggle to earn from streaming.

What kind of legal support does your platform entail? Do you have a team for that?

Yes. Here, artists can onboard their music on Hoopr, and the platform’s team helps with listing, tagging, and ensuring discoverability based on genre, lyrics, and occasions. Hoopr has a dedicated legal team that ensures agreements are in place for every song and use case, navigating India’s developing legal landscape where IP owners are increasingly sending legal lawsuits to brands for unauthorised usage.

So-What scale next?

Hoopr’s vision, for the next two to three years, is about focusing on further opening up the micro-sync licensing revenue stream by increasing market awareness among brands and creators. Our efforts will also involve helping users license and contextualise songs effectively, aiming to build a robust ecosystem that supports both creators and music partners.

pratimah@cybermedia.co.in