Recently, I ran into a longtime friend who works at a traditional media company. I hadn't seen him in a while. "So, I read somewhere that we're going to get disrupted," he said laughing, clapping a hand on my shoulder. "Have you ever thought about covering a different industry? Maybe retail? Or energy, or anything else?" Underneath every joke lies some truth. It's not pleasant being told (repeatedly) that your job, which is already hard, is going to get harder.
I deliver a pretty sobering message to traditional media companies. The current state of much of media is bleak, but that’s largely a lagging indicator of the disruption of media distribution over the last 20 years, enabled by the internet. I believe that another disruption is bearing down, the disruption of content creation, this time enabled by GenAI. Things will get tougher, not easier.
It’s not fun being the bearer of bad news. So, it’s fair to ask: why do it?
Foremost, I try to call it like I see it. But I also want to help people in the media business prepare for what's coming. I watched as TV was disrupted by Netflix. Disruption is a very powerful force and I doubt it could've been stopped. But, as an industry, we had a lot of own goals. There were things we could’ve done differently and, more important, a lot earlier. We might not have stopped the disruption, but maybe we could’ve blunted or contained it and, in some cases, capitalized on opportunities it presented.
I think that's also the case today, as long as traditional media companies confront the implications of GenAI head on.
Tl;dr:
GenAI will prove disruptive to media. But it isn’t yet clear how disruptive it will be. Disruption doesn’t always mean the same thing.
We don’t yet know how good the technology will get and to what degree consumers will embrace it, and for which use cases. It’s also up to incumbent media companies how they respond.
Another cause for optimism is that the declining costs of content creation will create opportunities too, and not just cost efficiencies. Whenever one input into a production process or value chain becomes more abundant, other things become scarcer and more valuable.
Here’s The Question™ that will determine how value will be created and redistributed in media over the next 20 years: as content becomes more abundant, what new scarcities will emerge, what existing scarcities will become more valuable, and what businesses (and business models) will be newly viable?
Here’s a partial list of emerging scarcities: compelling, truly original stories; curation (both algorithmic and editorial); human authenticity, provenance, and craft; tracking, monitoring, and measurement; consumer time and attention; marketing prowess and muscle; fandom and community; recognizable, trusted brands and IP; professional validation and infrastructure; nostalgia; talent relationships; and IRL experiences. Many of these are the sole province or in the wheelhouse of traditional media.
Confronting the potential implications of GenAI is overwhelming, but the worst thing traditional media companies can do is stick their heads in the sand. At a minimum, they should focus on workforce education and upskilling; move GenAI to the forefront of the long-range planning process; and develop skunkworks to build institutional expertise.
Disruption Comes in Different Flavors
Even if it stops progressing today, GenAI will be disruptive to media. It has all the hallmarks of a disruptive innovation, because it will dramatically lower the costs to create content, which will enable new entrants. Those new entrants will take some of the incumbents’ market share. That’s the definition of disruption.
But it’s important to understand that disruption doesn’t always mean the same thing. Even if GenAI will be disruptive, it isn’t yet clear how disruptive it will be.
Disruption doesn’t always mean the same thing. Sometimes it destroys incumbents, sometimes it doesn’t.
Clay Christensen, the father of disruptive innovation theory, didn’t explore why the effect of disruption can vary case-to-case. But consider Figure 1, which shows the stock charts of two “victims” of disruption. Digital photography disrupted Kodak; and the sharing economy (and specifically Airbnb) disrupted Marriott. Kodak went bankrupt. Marriott is near all-time highs.
Figure 1. Two “Victims” of Disruption
How Disruption Works
Why such different outcomes? Let’s think about this in the context of how disruption works and, in particular, the visual depiction in Figure 2.
Figure 2. Low-End and New-Market Disruption
Source: Adapted from The Innovator’s Solution.
There are two types of disruption, low-end and new-market. Most people are familiar with the former, which is represented by the far plane in Figure 2. Low-end disruption occurs when a disruptor enters a market with a less expensive, less performant (“crappier”) product. While this lower quality product initially falls short of most customers’ needs or expectations, it is “good enough” for the incumbents’ least demanding customers. So, the new entrant gets a foothold. Over time, its product gets better (i.e., it moves up the quality curve), closing the gap with the incumbents’ product, and picks off more of its customers.
Let’s take Airbnb. Initially, it offered customers a plainly inferior product—staying at someone’s house with limited privacy and no security, check-in desk, room service, or housekeeping. But it was good enough for some price sensitive hotel customers. Today, it offers everything from shared rooms to Airbnb Luxe, a curated list of properties that each come with a dedicated Trip Designer and the ability to arrange chefs, housekeeping, massages, and excursions.
New-market disruption is shown in the closer plane. The idea is that the upstart doesn’t only compete on the traditional attributes of performance, it also introduces new features. This closer plane is a new “value network.” If the upstart can draw customers (including customers who were previously priced out of the market) over to this new value network, that is new-market disruption.
Back to Airbnb. It introduced new features, like a full working kitchen, room to entertain, maybe a parking spot, a quaint neighborhood, outdoor space, privacy, etc. Customers who place a high value on those new features moved over to the new value network and they now have a different definition of quality in lodging. Many will only book Airbnbs.
A Framework for Thinking Through the Extent of Disruption
So, when we ask “what determines variation in the degree of disruption?,” what we are really asking, at least in the context of Figure 2, is this: 1) how far up the traditional performance curve will the upstart progress?; and 2) what proportion of the incumbents’ customers will it draw over to its new value network? Or, more simply: how big is the residual market that isn’t disrupted away?
Another way of asking “how disruptive is it?”: how big is the residual market that isn’t disrupted away?
We can develop a framework for thinking through these questions. I first wrote about this framework in How Will the “Disruption” of Hollywood Play Out?, but I’ve refined it since. The answer depends on:
1) How easy it is for the upstart to move up the quality curve. Sometimes, the new entrant can’t move all the way upmarket. Maybe it makes business model choices that foreclose the high end or it can’t overcome technological or regulatory hurdles. Airbnb will probably never successfully target business travelers.
2) How hard it is for the incumbents’ customers to defect. How sticky is the incumbent’s product? Customers usually have switching costs, like sunk investments in complementary products (your collection of Xbox games), the learning curve on the new product, and entrenched business relationships. Enterprise customers may be especially risk averse. Marriott and Hilton create loyalty programs partly to increase consumer switching costs.
3) To what degree—how much and how broadly—the upstart changes the consumer definition of quality. In a recent post (see Quality is a Serious Problem), I defined quality as the weighted set of attributes that consumers consider when making a purchase decision. When a new entrant offers new features that consumers value, it will change their quality algorithm. But how much and how broadly? In the case of Kodak, the benefits of digital photography (immediate feedback, zero cost of development, and massive storage), so outweighed the benefits of film that it utterly changed the definition of quality for almost all photographers. Everyone came over to the new value network! With Airbnb, this wasn’t the case. Leisure travelers are attracted to Airbnb’s new attributes of quality, but they still value the old attributes too. So, their definition of quality has shifted more subtly than in the digital film example. And business travelers’ definition of quality hasn’t changed at all. They don’t care about quaint neighborhoods or kitchens, they need the convenience, security, meeting space, and on site dining offered by hotels.
4) Whether the incumbent can blunt the threat. Incumbents may be unable or unwilling to respond to the competitive threat. “Unable or unwilling” rolls off the tongue, but they have very different meanings. Unwilling can be overcome. Unable cannot. Sometimes, they are unable to respond because of regulatory, legal, or contractual restrictions; maybe they are just too big or unwieldy to move fast; sometimes the upstart has a proprietary technology; maybe they don’t have the balance sheet flexibility or sufficiently accommodative investors to match the upstart’s pricing and reduce earnings in the short term; perhaps they have massive “technical debt,” or heavy sunk investments in physical infrastructure; or maybe the workforce doesn’t have the necessary skills.
The Die Isn’t Yet Cast
Now, let’s apply this framework to the disruptive threat of GenAI to media. I’d argue that #2 isn’t that relevant. Media customers don’t have high switching costs to consume different content. But the other three are highly relevant—and currently unknown.
We don’t yet know how far GenAI technology will move up the traditional performance curve. Will it become sufficiently realistic to pass modality specific Turing tests in video and music? Will it pass the uncanny valley? Will it be able to create sustained cohesive narratives or compositions? Will it be able to provide enough fine-grained creative control that professionals use it for the highest-level expressions of each format and medium?
We don’t yet know to what degree consumers will embrace it and, if so, for which use cases. Similarly, we don’t know if AI-native media (interactive, personalized, emergent stories, fan fiction, new formats, etc.) will prove compelling enough to draw a lot of consumers.
Lastly, we don’t know how incumbents will respond to the threat and opportunity of GenAI.
So, while GenAI will be disruptive, we don’t yet know how disruptive. And, importantly, incumbent media companies will have some influence over the outcome.
Revisiting What Becomes Scarce as Content Becomes Abundant
Another reason for some optimism is that abundant content will create opportunities, for three reasons:
New abundances create new scarcities (or moats). It is a foundational economic principle that value flows toward scarcity. Abundant things are commoditized and scarce things command a premium. When an input into the production process or part of the value chain becomes abundant, other things become the new chokepoints and increase in value. In media, for instance, lower barriers to distribute content resulted in an explosion of content choices. That made curation scarcer, shifting value to the platforms that control the end user relationship.
New abundances amplify existing scarcities. A good example is live music. Live events have always been scarce. Only so many people can fit into a venue. But as recorded music approached free, live concerts became relatively scarcer, more valuable and, as a result, more expensive. This happens because resources can be reallocated from the newly abundant thing to the already-scarce thing. (Money no longer spent on physical music could be reallocated to going to shows.) Also, the social signalling of the scarce thing goes up as the alternative becomes commoditized, especially in our increasingly social world. People used to line up outside Tower Records to be the first to buy a new release. If anyone can listen to the new Beyonce album when it drops, being at the Beyonce concert is even more valuable.
New abundances enable new businesses. When an input into the production process gets much cheaper, that makes new businesses economically viable. As an example, bandwidth used to be extremely expensive. The emergence of essentially free bandwidth created literally trillions of dollars of value because it birthed streaming media, cloud computing, and SaaS business models, among other things.
The challenge is to think through this question: as AI-enabled content becomes abundant, what new scarcities will emerge, what existing scarcities will become more valuable, and what businesses will be newly viable? I’ve taken swags at this before, but I think it’s worth continually revisiting this question. In a lot of ways, it is The Question™ that will determine how value will be created and redistributed in media over the next 20 years.
Here is The Question™ that will determine how value will be created and redistributed in media over the next 20 years: as content becomes abundant, what new scarcities will emerge, what existing scarcities will become more valuable, and what businesses will be newly viable?
When trying to answer it, there are a couple of helpful lenses to use. One is the perspective of consumer decision making. When confronted with infinite choice, what filters will consumers use to cut through the clutter? Another is the consumer definition of quality. How do consumers define quality in media? Which quality attributes can be satisfied by GenAI-enabled content and which can’t? Yet another is the needs of creators. If anyone can create, how will creators distinguish themselves?
Here’s what we can be pretty confident will remain scarce or become relatively scarcer:
Compelling, Original Stories
Toward the end of the opening credits of the 1992 movie The Player, a dark comedy satirizing Hollywood, Robert Altman sets the tone by allowing the viewer to eavesdrop on a few writers’ pitches. One project is likened to The Gods Must be Crazy meets Pretty Woman. Another is compared to Ghost meets The Manchurian Candidate. The scene skewers Hollywood’s tendency to frame everything as derivative.
Derivative will be commoditized. Novel will not.
GenAI excels at derivative: movies, songs, and games that are kind of like something else. In the future, derivative will be commoditized. Truly novel ideas, formats, structures, and combinations—ideas that feel intentionally novel, not randomly thrown together—will be ever more valuable.
Curation, UI, and First Party Data
This is a simple extrapolation of the current state. Today, consumers can’t possibly navigate all the content that exists. Creators upload more than 500 hours of video to YouTube every minute. Roughly 100,000 new tracks were uploaded to DSPs every day last year, according to Luminate. There were 19,000 new games released on Steam last year too. When you add in Reels; posts on X/Twitter, Facebook, and LinkedIn; and news articles, it is overwhelming.
So, curation is more valuable than ever and it will get only more so as the volume of content increases. That benefits the platforms, of course, which have become the new distribution chokepoints. They control the user interface and the ability to influence what customers consume next. They also extract vast amounts of first-party data which, in turn, they use to improve the quality of their recommendations.
Owing to the volume of content they host, most platforms lean heavily or entirely on algorithmic curation. How they extract value from this curation depends on their business models. YouTube is (mostly) advertising supported and owns (essentially) no content, so its chief goal is to serve recommendations for videos (and auto-play subsequent videos) that are more monetizable (have more ads per hour) and keep viewers on the platform longest. Spotify has a different model, because it is mostly subscription based and must pay for content. It also wants to increase listener satisfaction and listening time, both to sell more ads and reduce churn. However, because it pays more for some content than others, it concedes that “…commercial considerations, such as the cost of content or whether we can monetize it, may influence our recommendations.”
Editorial curation is also growing in value. Consumers are increasingly turning to trusted brands and influencers to help them navigate the tsunami of options. This is, in effect, how most influencers create value. Many newsletters have taken on the same role, such as The Skimm and Morning Brew.
Fewer traditional media companies have seized this opportunity. The New York Times is an exception—Wirecutter and Cooking are both trusted curators—as is Conde Nast with Vogue, GQ and Bon Appetit. It’s one thing to recommend products or curate recipes, it’s quite another for a media company to recommend competing content. Even so, the value of trusted editorial curation will only go up.
Human Authenticity, Provenance, and Craft
The biggest trend in watches is independent watchmaking. Not Rolex or Patek Philippe, but non-household names like Philippe Dufour, F.P. Journe, and Greubel Forsey. Their rising popularity reflects a rejection of mass production techniques and a rising premium for the artisanal and handmade.
That will probably happen with content too. As it becomes easier to mass produce synthetic content, the best songs, movies, TV shows, and social posts that convey authentic humanity will likely be more valuable than ever.
The connection with the artist or creator will be more important than ever, especially with content for which provenance and backstory is an element of the experience.
Also, for some content, the connection with the artists and the backstory behind the content are critical attributes of quality.
Entertainment media is mostly an emotional good. People choose a film, novel, show, song, or game because they want to feel something. There are many cases in which the ability of a piece of content to evoke emotion is inextricably tied to the provenance and backstory behind it. You can’t and don’t want to separate Taylor Swift’s relationship history from a Taylor Swift song; Jay Z’s upbringing in the Marcy Houses from a Jay Z song; the tension between Mr. Beast’s earnestness and sometimes brutal challenges from a Mr. Beast video; Quentin Tarantino’s unapologetic geekiness about film from a Tarantino movie; Hunter S. Thompson’s drug use from a Hunter S. Thompson story; or Picasso’s artistic evolution from a Picasso. When you watch A Complete Unknown, your knowledge of Timothee Chalamet’s five-year effort to learn guitar and sing is a critical part of the experience.
By definition, synthetic content can’t replace humans in any context in which human authenticity, drama, provenance, craft, or lived experience is a critical attribute. To paraphrase something Chris Dixon has said, computers are now much better than humans at chess, but watching humans play chess is more popular than ever, while no one wants to watch two computers play each other at chess.
Tracking, Monitoring, and Measurement
Cross platform media measurement is already hard. As content proliferates, including AI-created and derivative content—across who knows how many platforms—it will become ever more important to track:
Provenance - who made it?
Rights - is it authorized?
Distribution and monetization - where is it consumed and, if applicable, monetized?
Consumption - who’s consuming it?
Measurement - how is it performing?
Attribution - did it lead to a desired outcome?
There are a variety of technical approaches to this (watermarking, metadata tagging, fingerprinting, automatic content recognition), but no scalable, interoperable system covers all content types and platforms. Plus, many platforms, like YouTube and Facebook, will likely continue to restrict data sharing, making this even harder. Maybe it’s an unsolvable problem, but it could be massively valuable if anyone can figure it out.
Consumer Time and Attention
The whole point of media is to attract consumer attention, so it probably seems obvious that attention will be more valuable as the volume of content increases.
Traditional media companies still command a lot of consumer time.
What I mean is that traditional media companies, even if in decline, still command a lot of attention. For all the discussions about linear TV dying, for instance, in the U.S. it still represented 60% of all viewing hours last year, according to Nielsen. According to Edison Research, about 1/3 of time listening to music in the U.S. is on AM/FM radio receivers, more than streaming services or YouTube.
The time spent on traditional media declines every year, but this is still very valuable real estate that will probably be more valuable in a world of infinite choice.
Marketing Prowess and Muscle
The marketing function has gotten a lot more complex over the last decade or so, owing to the proliferation of media channels, the emergence of sophisticated targeting and measurement tools, the advent of social, and the speed of change.
In media and entertainment, most marketing is for specific shows, movies, albums, games, or books. It used to be that great marketing equated to good creative, maybe coupled with good PR. With fewer media outlets, media plans were pretty formulaic. The creative itself was derivative of the content—you cut a trailer or two; developed an ad for the newspaper, billboards and bus kiosks; and a radio spot. The timing of the campaign was also pretty formulaic, timed to the release of the movie, game, TV show, or album. Once set in motion, it was just a matter of execution and monitoring.
Marketing is far more complex and important than it used to be. In a world of infinite choice, it will be even more of both.
Today, great marketing requires better, more resonant creative, because it is harder to cut through the noise. It also requires far more sophisticated optimization of paid media spend through media mix optimization; far more complex measurement across multiple channels; acquisition and utilization of first-party and third-party data for targeting and measurement; creating attribution loops; A/B testing of creative; and perhaps even personalization.
Because of social, earned media is potentially vastly more valuable than a few positive articles. Great marketers can multiply the impact of their paid spend many times over by building up communities, picking the right influencers, and creating content that encourages sharing and organic virality. But they must also be highly attuned to these organic signals arising from the network (which they don’t control and may deviate from what they intended or hoped) and adapt quickly.
As the volume of content increases, great marketing will become even harder and, for that reason, even more valuable.
Fandom and Community
I’ve written about fandoms a lot, like IP as Platform or here, where I described what I call “fanchise management.” So I won’t belabor the point.
All media is social. Fandoms provide a sense of connection, a common vernacular and sometimes even a shared value system. When faced with infinite choice, one filter consumers will likely use is the appeal of the fandom and community surrounding an intellectual property. So, committed, active, and large fandoms will also be increasingly valuable.
Fandoms will become an even more important filter as choice increases.
In many cases, media companies don’t do much to support or foster fandoms. They usually exist off platform, like on Reddit, TikTok, or Discord. In the future, I think IP owners will work much more concertedly to develop, support, and manage fandoms. Something I keep harping on: as GenAI brings down the cost of creation, the most progressive IP owners will encourage fans to remix and reimagine their favorite IP to create even stronger bonds.
Recognizable, Trusted IP and Brands
Another filter consumers will probably use to cut through the clutter is familiarity, especially when they have a pre-existing positive perception of that brand or IP. This is partly due to what behavioral economists call the “mere exposure effect:” people tend to like something just because they’ve been exposed to it before.
Importantly, this doesn’t only mean Harry Potter, The Fast and the Furious, or The Legend of Zelda, it also means brands as signals of quality, whether Netflix, Warner Music, or Rockstar Games.
Professional Infrastructure and Validation
One of the tectonic themes in media I write about is disintermediation. Historically, the most powerful participants in the value chain—and the household names—have been the intermediaries between creatives and consumers: studios, labels, and publishers. As technology democratizes each step of the media product development process—production, marketing, distribution, and monetization—this shifts bargaining power to creatives or enables creators to circumvent middlemen altogether and go direct to consumer.
As the cost of creation and distribution falls, the value of professional validation will also rise.
But just because creatives can circumvent the traditional intermediaries doesn’t mean all will want to. Many still want the endorsement of a recognized brand and the professional infrastructure they can provide, such as financing, legal and rights management, data analytics, marketing muscle, knowledge of international markets, and access to talent networks. This is, after all, what is happening in music. Today, almost all new artists emerge from the tail of self distribution (Soundcloud, TikTok, YouTube). But once they have a hit, the first thing they do is sign a major label deal. Similarly, in book publishing, authors can easily self publish and will retain a much higher revenue share, but most still want the “Good Housekeeping Seal” of approval of working with an established publisher.
As the cost of creation and distribution falls, the scarcity of professional validation will increase. Since studios, labels, and publishers have limited capacity to support talent, the signalling value of working with a “major” will go up.
Nostalgia
As life gets more complex, there is a growing demand for nostalgia. Current examples include rising popularity of vinyl records, vintage shopping, retro design, 8-bit/16-bit video games, and even a resurgence of film camera sales.
Nostalgia has value because of its association with the past. No matter how good GenAI gets, it will never be able to go back in time (I think).
Talent Relationships
Even if GenAI proves to be a democratizing technology and lowers the barriers for tens or hundreds of millions of creators, there will always be a finite number of truly talented people. There will be even fewer talented people with an established reputation and brand equity.
In the future, as the barriers to creation fall, these people will likely have more choice how and with whom to make content. Having strong relationships with that talent will also grow in value.
IRL Experiences
This is an extension of what I wrote about music before: as information goods get cheaper, experiences become relatively more valuable.
As content becomes even more abundant, this divergence will probably get more extreme. I have seen suggestions that the key to reviving the movies is to lower ticket prices. The right answer may be the opposite: make going to the movies an even more rarefied, special—and expensive—experience.
What Big Media Companies Should Do
What should jump out from the prior section is how many of these elements are accessible to media companies: compelling stories, editorial curation, human authenticity and provenance, marketing prowess, fandom and community, recognizable IP and trusted brands, professional validation, nostalgia, talent relationships, and IRL experiences are all either the sole province of traditional media or at least in their wheelhouses.
Almost every company is currently working to “figure out” AI. They’re trying to determine their data preparedness to support AI applications; figure out the best operating model for AI deployment and governance; navigate the legal issues; decide whether to rent, build, or buy AI capabilities; and prioritize use cases.
If I were running a traditional media company, here are a few other things I’d do:
Education and Upskilling
Some media CEOs have been publicly skeptical or dismissive about GenAI and its potential to affect their business. That’s the wrong message to send.
Everyone in the organization should understand the potential for GenAI to change how they work—not just the software engineers, customer support, and marketing departments. One way to deliver that message is to ensure that everyone is trained on GenAI: how it works and how to use it.
Planning and Resource Allocation
The potential effects of GenAI should be incorporated into the planning process for every media company, especially the long-range plan (LRP). How might it affect the core business? What opportunities might it surface? Since so much about the evolution of GenAI is still unknown, one very helpful tool is scenario planning.
In a recent post (How Far Will AI Video Go?), I walked through an example of how this works, a scenario analysis for AI video. I created a 2 x 2 matrix by varying two key drivers (technology development and consumer acceptance) between their logical extremes, to produce four scenarios: low tech development, low consumer acceptance (“Novelty and Niche”); high tech development, low consumer acceptance (“The Wary Consumer”); low tech development, high consumer acceptance (“Stuck in the Valley”); and high tech development, high consumer acceptance (“Hollywood Horror Show”). I then wrote out a narrative for each scenario.
The goal is to understand the range of potential outcomes and the appropriate strategies in each—including what dominant or “no regrets” strategies would work in all scenarios.
I think every media company should do this, because it will inevitably yield insights. In the case of the exercise I did for video, here are some of the imperatives that jump out:
Rethink the content development process. In the future, the concepts of “TV shows” and “movies” (the distinction between which has already blurred as TV production values, budgets, and episode duration have all climbed) may give way to “narrative video” — which might be short, medium, or long duration and might be episodic or self contained. Whatever you call it, there will likely be a lot more of it. Since demand for video is more or less fixed, success rates will fall, probably by orders of magnitude. Back in the days of broadcast, for every 100 shows that went into development, 10 might get pilot orders and one would get a series order—a hit rate of 100:1. In the future, the hit rate might be 1,000, 10,000, or 100,000:1. How will development work then? It would have to look completely different. The locus of development will probably shift from content creation to discovery, rapid experimentation, iteration, and co-creation. While there might still be a high touch, manual selection and development process for some very high-budget shows and movies, for most content it will probably shift to more of a data analytics job: scanning the network algorithmically to surface potential hits; cheaply and quickly prototyping story ideas; running A/B/C testing on storylines; allowing fans to influence story development, etc. This would obviously be a big change from how development works today and people who think of TV as an art form will hate it. But there’s no way around it: if hit rates plummet, it will require doing things differently. The scenario analysis requires that you confront the question.
What happens if the hit rate on “TV shows” goes from 100:1 to 10,000:1?
Be intentional about content genres. Another implication is that content that is more dependent on authentic human connection is more insulated from AI-enabled competition than content that isn’t. If you’re in the unscripted business, for instance, formats that rely on human drama—singing and cooking contests, dating shows, game shows—will be far more insulated than formats that don’t, like true crime or historical docudramas. Also, a critical technical hurdle for AI video is the uncanny valley. So, content that doesn’t rely on emotive human faces is also more susceptible to competition from AI-enabled content. Studios should be particularly careful about allocating a lot of resources to animation, which is at risk of being commoditized.
Certain genres will be more insulated from GenAI than others.
Invest in fandoms and fan creation. For the reasons described above, IP owners should be far more proactive developing and managing fan communities—not leaving it to chance. As I mentioned, I think this should include encouraging fan creation.
Figure out how to participate in the growth of the creator economy. GenAI will likely close the gap in production values between corporate and creator content. As a result, it will empower creators more than it will empower established studios and likely accelerate the growth of the creator economy. It will raise the stakes for traditional studios to develop a creator economy strategy.
Lean into provenance and backstory for all content. The creators and stars behind all TV shows and movies (and, for that matter, all media) should be humanized whenever possible. This means more than just press junkets, but working harder to communicate the backstory and process of making the content. Music is already great at this, because the artist and the art are inextricably linked. TV, film, and games, not as much.
Invest in IRL events. Every copyright owner should be figuring out how to develop live or otherwise scarce experiences around their IP.
Rethink media marketing. Above, I mentioned that marketing has become much more complex in recent years. In a world of abundant content, this just increases. Marketers will increasingly need to think about marketing more as an ongoing system—hyper-targeted or even personalized creative; continuously iterated; constantly monitored—than as a campaign.
Envision what new business opportunities might arise. When content creation becomes cheap, what becomes newly possible or a new bottleneck? Rights and authenticity tracking and verification? Algorithmic content listening? Personal AI agents that sift through content to surface what’s personally and contextually relevant? Fans clubs as a service?
Create a Skunkworks
A lot of media companies are reluctant to embrace GenAI. Understandably, they’re worried about talent backlash or unresolved legal issues. At most, some are experimenting here and there with weaving AI into their existing workflows.
The risk of waiting for all issues around GenAI to be resolved before experimenting is being caught flat footed. And trying to adapt AI to existing workflows will yield fewer benefits than changing workflows to adapt to AI.
This approach creates two problems: 1) they might be caught flat footed when and if the technology becomes more widely accepted and legal issues are clarified; and 2) trying to adapt AI to work with existing workflows is likely to yield fewer benefits than trying to optimize workflows to work with AI. Every company has “org structure debt,” “process debt,” and “workflow debt”: “this is how we do it, and these are the people that do it.” Adapting AI to fit “this is how we do it” will likely be a lot less illuminating than starting with a clean piece of paper.
One solution is to run a skunkworks project: create internal teams, give them resources (money and perhaps even access to fallow IP), and see what they can create. Best case scenario, they create something worthwhile. Worst case scenario is that you learn something and develop the internal know-how to deploy GenAI when it becomes necessary or at least more palatable.
Source: Tom Toro, via The New Yorker
Realistic Doesn’t Mean Fatalistic or Nihilistic
AI is going to turn a lot of industries upside down. The compounding problem for media is that the industry is already struggling with the effects of the last disruption. It’s easy to feel overwhelmed.
The good news is that traditional media companies have unique brands, assets, and capabilities that will become more valuable as content becomes more abundant. But to capitalize will require concerted effort and, at the least, confronting tough questions head on. Wishing things away is rarely a good strategy.
I think AI's true foothold in the media industry will continue to be what we're seeing the in the rest of the corporate world - a little bit for creative endeavors, but mostly as a workhorse for boring, repetitive tasks performed in the background. And really, aren't those tasks the major barriers to entry anyway?
And on your prediction about controlled communities becoming the new marketing campaign, I would say it's already come true. Nearly every Reddit community I've seen for a major television show is controlled by somebody affiliated with the owner/distributor.
Great post, thanks for sharing.
Great article.
1. Disruption isn't a force in itself. It is happens because someone finds a better idea and the incumbents keep trying to sit on their success or trying to avoid being disrupted.
2. Airbnb vs Marriott was a clever piece of marketing by Airbnb to position themselves outside the short term rental industry. They never really were a big threat to Marriott or the hotel chains. Airbnb's real competitors are Booking, Expedia and so forth. But they made everyone believe they were a hotel company, which was a smart move.
3. Funny quote "It’s not fun being the bearer of bad news. So, it’s fair to ask: why do it?" - literally 99.9% of what news media is, is being a bearer of bad news.
4. IMO: Because long form articles will be so easy to make with AI. Short form with thrive. Not sure that's a great thing.
5. Trusted people will win. But there will need to be some kind of curation. It can be by AI. Subscribing to dozens of paid substacks isn't viable. We're in an unboundling phase, we're soon going back into a re-bundling phase.