Media — Impact of Gen AI
Media serves as a conduit for storytelling — whether it is a movie, a song, an image, an essay, a TV show or even a game. Over the past few decades, the internet democratized access and distribution of all forms of media. The next frontier is Generative AI which will empower individuals to tell professional level stories. Just as chips reduced the cost of compute, web reduced cost of distribution, Gen AI will reduce cost of creation.
I wanted to spend some time thinking about the potential value creation / disruption opportunity in this space.
How does Gen AI change the current movie / media making process:
- Concept development: Gen AI can help reduce time/ cost by 20–40% through more efficient script-writing. Journalists are already using Gen AI to simplify research. TV producers are using it to edit storylines. DuoLingo is creating content for its language learning app. In gaming, developers are building game characters and dialogues for NPCs. In social media, people are creating avatars!
- Pre-production: Limited impact here. An assistant director I spoke with mentioned that Gen AI can help automate scheduling for cast and production members which currently takes up a lot of his time.
- Production: Gen AI’s role is again somewhat limited. But for generic scenes, such as a Manhattan skyline or a forest storm, AI comes handy. Btw, Show-1 is completely mind blowing.
- Post-production: Our discussions indicate that post-production time / cost can potentially be reduced by 20–40% (see post product process below). Editing becomes easier, it can help reduce human hours, remove unwanted objects. You can also do automated dubbing, make actors look younger etc. One short film producer told me that time to edit has gone from months to weeks. Game developers are saying the same. Another application is catching dead pixel or the Starbucks on Game of Thrones. AI can also mimic / change a celebrity’s voice
One Hollywood director I spoke with sees Gen AI potentially shaving 20% off movie production costs soon which can reduce movie budgets back to 2000s levels (see chart below).
For creators/ indie developers who release short-movies on YouTube -> Gen AI can help with 50% reduction in production costs -> which means their profit dollars on reaching 1M viewers can reach ~$15k (based on $20 CPMs, 3 ads shown for a 15 min video).
Areas of disruption/ potential impact
Currently, there is $2T of market cap among leading media companies, but tiny wealth/ value generated by creators. As Jensen said we are going from rendering pixel to generate pixels in real time which will empower us to tell stories that are much more descriptive and powerful. As this happens — there will be tons of value creation opportunities:
- Shift of value / power from labels/ media companies to creators: Labels, production houses and even CPG brands typically capture most of the value that an artist creates for them. As cost of content creation / production declines — creators can own/ distribute their own content.
- Explosion of semi-professional content: Average movie production budget currently is $50M which can possibly decline by 20%, resulting in number of movies / TV shows released in US to grow from 500 / 2500 to 3,000 / 20,000+
- Celebrities can own / license their content — and developers can build on top. Elon could be your personal therapist, Federer could be your personal tennis coach on Nike app etc. Messi can be your personal cheer leader. Somewhat spooky but creators can become immortal.
- Personalized content: AI can generate personalized stories and essays (fiction and non fiction) for users depending upon their mood or requests. Personalized advertisements and training videos can be made as well
- Enterprises can use Gen AI to offer personalized training and on-boarding
Traditional media businesses that made 30%+ EBIT margins will face headwinds as content creation gets democratized and churn increases. They were already under pressure as twitter, youtube and reels democratized distribution and offered higher share of ad revenue.
