Ex-Napster/Condé Nast exec, Miso CEO, CLI expert on AI and content licensing

I am intrigued about how AI interfacing may change publisher models. Suppose your initial engagement starts not with a branded newsletter, website or app, but with a simple query to a Large Language Model like ChatGPT, e.g., “Give me a summary of the main tech headlines from the past 24 hours?”

This completely upends the link-based model with branded gateways into content experiences. Why should we then package data (content) into a branded wrapper at all? Especially if those LLMs pay content licensing fees for that data.

But then you have a conundrum. It’s really only big media that have the leverage to get an OpenAI to a table to talk and do a deal (e.g. Axel Springer) or walk away and sue for more (e.g. The New York Times). Does this then open an opportunity for content licensing specialists to play a more active role in representing a conglomerate of smaller, independent publishers? Or should we be looking towards a Spotify-type model, where content creators get paid by play, however flawed that model may be?

Technically speaking, as one of the experts below also explains, this is already possible…

In this article

1. Content licensing in the AI age.
2. Why publishers must avoid one-time payments.
3. Using technology to identify (and pay) content sources.
4. Insight from a former Napster associate on not repeating past mistakes.
5. A technologist’s view on “royalties and residuals” as a model for payments.
6. A licensing expert on why publishers must be proactive, or they’ll miss opportunities.
7. Reflecting the 1990s and the need for adaptability and experimentation.

I asked two content licensing experts and an AI technologist for their thoughts.

I’ll start with Angie Byun because of all the Napster mentions we see in relation to where the publishing industry is now. Angie was an associate at Napster for a short while. While there remain unanswered questions regarding international content licensing, she urges publishers “not to repeat Napster-era mistakes”.

I asked Miso Technologies’ Lucky Gunasekara for his thoughts. At Miso, he works on combining small data and Large Language Models to create high-performance search and discovery experiences for its partners around the world. At our Mx3 AI event in London in December, he was one of the audience favourites. See an off-stage interview with him from the event here. Below, Lucky explains why he thinks “one-time” payments are a “bit of a trojan horse”, why “royalties and residuals” provide an alternative model, and, critically, why identifying (and paying) for AI content sources is already possible.

Finally, Frank Bilotto, Business and Legal Affairs, Creative Licensing International in the US, brought licensing, legal and M&A chops to the conversation. As a former Legal Counsel for Content Licensing at Thomson Reuters, Frank brings a wealth of experience that extends beyond legal guidance and includes playing integral roles in the launch, growth and acquisition of three digital content-related companies totalling almost $500 million. He warns against accepting one-time payments for content but says publishers must be proactive not to miss out on the AI opportunity.

Angie Byun: From Napster to Today

Angie Byun is the US-based principal and founder of AB World and was formerly an international brand and content licensing senior exec at Condé Nast, where she worked with brands such as Vogue, GQ, Golf Digest, Wired, The New Yorker, and Pitchfork (RIP).

She also worked at Napster – the peer-to-peer MP3 file-sharing platform – for a brief period in 2001. Napster was soon in music labels’ and artists’ crosshairs, and after multiple lawsuits, Napster as we knew it was gone. The technology that enabled it was not, so there was no return to the past and adapting to new business models was the only option (here’s a quick Napster overview).

I asked Angie about her Napster days and AB WORLD, where, among others, she negotiates content licensing deals around the world. In particular, I asked if there were any parallels that she could draw upon from her days there and subsequently in publishing, and how she thinks things like content licensing might play out in the AI age.

“One of my first jobs out of law school back in 2001 was as a compliance associate at Napster, which at its peak reached 80 million users. My tenure there was short-lived as it coincided with the company’s legal battles against the major record labels, various music artists and the Recording Industry Association of America (RIAA), all suing it for piracy and copyright infringement. 

“Napster’s eventual demise didn’t result in more power and authority going back to the record labels and artists, but merely served as a catalyst for other music streaming platforms and business models like Spotify and Apple Music to emerge – fundamentally altering how music is distributed and how artists are compensated for their work.   

“In hindsight, I believe many artists and record labels regret their adversarial stance towards Napster, recognising the missed opportunity to collaborate and capitalise on the seismic shift that was happening in the music industry.”

Angie says she is struck by the parallels between the Napster days and now with the adoption rate of LLMs such as ChatGPT (100 million weekly users) and the impact it has/will have on publishers.

“We see how media organisations and news outlets grapple with copyright infringement issues associated with the ‘unauthorised’ use of their content on these platforms. From the New York Times lawsuit to prominent authors pursuing legal action against OpenAI, it begs the question of whether these entities should view these platforms as potential allies rather than adversaries.  

“By working with and not against the LLMs, might there be a larger, strategic opportunity for media companies to monetise their new and archival content, expand their audience, and innovate their advertising-dependent revenue models?  And while LLM’s are in its nascent stage, can both parties develop creative solutions to properly compensate, cite and regulate content that is in line with ethical and legal standards?”

It is perhaps too soon for me to ask how all of this may play out in the content licensing space, where I believe the need for access to quality content may continue to demand a premium while the cost of other, more generalist content plummets.

“The potential impact on the licensing, marketing and distribution of content internationally is an interesting consideration. My consultancy, AB WORLD, is in the business of licensing and marketing Intellectual Property into international markets under license. I am curious how this will affect the premium licensing fees companies have traditionally commanded for their IP.”

One of the reasons is that, as with other big technological advances, the landscape will likely shift considerably in ways we may not foresee now, while power balances may shift when traditional “constraints” are democratised. As Angie points out, LLMs may enable native-language publishers to reach new audiences in another language.

“With LLMs, media brands may find themselves capable of producing their own international editions, freeing them from the reliance imposed by traditional publishing partners. Conversely, international publishers may find that they can produce more global content that is readily available and near-perfectly translated [strengthening them further].

“It will be interesting to see how the business of international licensing will be impacted, but one thing I know for sure is that AI is here to stay. Taking lessons from Napster, we need to seize the opportunity to adapt to the changing paradigm, rather than repeating the missteps of the past.”

Lucky Gunasekara: ‘AI can identify sources and pay accordingly’

Lucky Gunasekara is the co-founder and CEO of Miso Technologies. I asked for his views via email.

“Yeah, I’ve been thinking about the same topic,” he replied, “and am trying to find the time to write a blog post on ‘Royalties and Residuals, Not One-Time Payments’ as I’ve been piecing my notes together.

“In general, I think the idea of one-time licensing is actually a bit of a Trojan horse by AI companies. Yes, they are paying newsrooms and publishers for access to their content for training. But unless they link to their sources in their outputs and pay on a ‘per Answer’ basis, like a streaming royalty or a TV/Film residual, this is just a one-time payment for them and a massive loss for the industry.

“Hollywood figured out a long time ago that the right way to pay artists for their work being redistributed was to pay continuous royalties and residuals, and there’s an entire world of digital forensics now in streaming to ensure this in music. The SAG and WGA strikes took a harder look at this covenant as streaming sites, and studios were baulking at this model in the world of film/TV streaming.

“I think media and news need to get to the same conclusion on how to have AI compensate their work. And this is very, very possible. We’ve proven this with our own build of Answers [at Miso] that it’s very doable to identify which sources fed an Answer and, therefore, pay those sources a fair cut of any revenue that Answer generates.”

Frank Bilotto: ‘Avoid one-time payments for content’

Frank heads up Business and Legal Affairs and Creative Licensing International. He responded to me via email.

“Before publishers recognised the value of digital content, I negotiated content licenses for the world’s largest aggregator, when publishers were happy to accept 5% royalties. In those early years, publishers were short-sighted, as they believed revenue from digital was irrelevant, and all the real money was in print. It took 15 years for publishers to finally understand the value of their content in digital form, and for royalty rates to reflect that value.

“Today, publishers look at AI with a similar myopia. This isn’t complicated. AI will change the value of content as we know it. Unless publishers properly position their content within AI models, they will miss the opportunities to partner with AI developers to expand their reach and revenue.  

“We’re not advising clients to go all in right now. CLI is still cautious about ChatGTP, as they have the mistaken idea that if the content is available on the internet, it is free to be used in any manner. The practice of AI technologies confiscating revenue and branding opportunities from publishers is nothing new. Over a decade ago, Google’s early AI started to provide answers in the first search results, instead of hyperlinks to the publisher’s website. Google’s promise not to be evil has been eviscerated as publishers continue to lose countless opportunities to monetise the content they create because users can see the information they need without ever leaving Google.

“So, publishers are justified to fear AI’s current exploitation of content. However, we advise our clients to consider some select options to partner with AI developers, two of which I will detail here. 

“First, major aggregators are using AI to improve the discovery capabilities of search. We are advising all our publishers to opt into these agreements. Properly implemented, AI will make smaller publishers with premium content more competitive against larger brands, as content that once was buried in a cemetery of small publishers, will be resurrected for all to see.

“Second, we’re advising our clients to avoid one-time payments for their entire archive of content unless the offer is so big that it significantly changes their balance sheet. We do favour long-term subscription models that provide a significant upfront fee for the archives, followed by annual fees for new content that reflects the full shelf life of the content. We’re protecting our clients against the risk of termination of subscriptions, by including provisions such as continued royalty payments post termination to fairly compensate publishers. 

“AI companies will have to develop technology that can accurately identify the content source used to generate artificially [re-]created content. CLI has been an ardent advocate for publishers to demand that AI companies develop this feature, which is not terribly difficult to do [CH: correct, this is already possible – see Lucky’s comments above]. When this happens, and it will, publishers will be able to enter agreements with AI developers with confidence that they will be fairly paid for the use of their content. The first AI developers to successfully implement this technology will lead the marketplace in attracting premium publishers to their generative AI technologies. 

“Unfortunately, if the past is prologue, most premium publishers will bury their heads in the sand and repeat the same mistakes they made in the early days of digital content. For those that recognise the value of their content to AI developers and structure smartly drafted agreements, AI can provide a windfall to publishers similar to the one that digital content revolution has.”

Party like it’s the 1990s!

After bubbling underneath the surface for some years, the sudden “mass” adoption of AI feels a bit like the start of the consumer Internet era in the 1990s. I was a relatively young journalist at South Africa’s Naspers Media24 division at the time. The group launched News24.com under the direction of Arrie Rossouw, who had recently spent three years in Washington DC as the group’s newspaper correspondent in the capital.

It was through Arrie’s first-hand AOL-era Internet experiences in the US and a visionary leader like Koos Bekker heading up Naspers (who soon after acquired a stake in Tencent in China, which turned out to be one of the best Internet startup investment stories) that News24 came to life.

Media24 was quite progressive towards News24, allowing us to use content from the group’s print titles to add uniqueness to our more vanilla offering from the news wires (i.e., licensed content). The newspapers were predominantly Afrikaans, so we would take the stories, translate them, and publish them in English.

Most newspaper editors were, for obvious reasons (free versus paid content), not completely enthralled with the idea – sometimes threatening to cut us off their content feeds if we did not behave (which mainly meant not publishing certain stories before an agreed time after the newspapers had been out in the market for a few hours).

We were encouraged to experiment, including trying out revenue streams at News24 in addition to display ads.

In 2001, after the dotcom bubble burst, we (News24 in South Africa) took a first foray into a paid content model**, targeting “overseas” users based on IP addresses. It was crude, but you experiment and learn. This was later folded into a wider group offering called the Kudu Club, which included, from memory, news and other content from the group for South African expats. I don’t think it lasted very long, but it has morphed into new paid offerings that are doing much better today than back then.

In another experiment, we licensed the translated content to the group’s SuperSport (a pay-tv sports channel on the group’s satellite TV platform DStv) for what was, for us, quite a handy income stream! In another experiment, we launched a News24 channel on DStv, a text offering on your TV, basically mirroring the website, giving us another handy income stream. In this “new” world, we could create stuff once, expand its distribution, and make money off it more than once. Happy days.

In any event, I learned a few things during that time, such as:

1. When a new technology threatens established business models, there will always be friction. It’s just how it is.

2. Technology wins. If something comes along that is more efficient than what had gone before, it will win through sheer adoption by consumers. You must accept it, lean into it and see how you can make it work for you. Sometimes something will. Sometimes nothing will. But there is no more “business as usual”.

3. You must experiment. Sure, we messed up, but it wouldn’t have changed the future of the group’s newspapers one iota if News24 was prevented from publishing their stories.

4. Existing business models will be destroyed. New ones may be vastly different from the older ones. Media was glamourous BIG business. Now it’s just business. There’s nothing wrong with that.

Exciting times.

** News24 is still alive and kicking, and much better than us back in the day at selling premium access!

Join us

  • Join us in person at our Mx3 Barcelona innovation summit on 12-13 March, where we will discuss several themes dominating the current and newly emerging media environment. It’s immersive, off-the-record, and down-to-earth, with places strictly limited in line with Dunbar’s 150 ;). See more at Mx3barcelona.com. Get a FIPP discount with code MX3_FIPP10.
  • On 4-6 June, we are in Cascais, Portugal for the 46th FIPP World Media Congress, which we organise and host on behalf of the international media body FIPP. Learn more at fippcongress.com.
  • There is more on their way soon: check back in at mediamakersmeet.com or sign up for the Media Makers Meet newsletter to get updated.

This article was originally published on the mediamakersmeet website, here.


Your first step to joining FIPP's global community of media leaders

Sign up to FIPP World x