SoulScape Day 1

Notes from a day of panels on AI, storytelling, and the future of creative IP.

Panel: The New IP Economy

Panelists: Mark Holmes, Kabir Gill, Wen Song

What is the value of IP?

IP protection is becoming increasingly critical as AI blurs the lines of creative ownership. ElevenLabs, for example, requires explicit permission before using a performer's voice. The core of IP isn't just legal protection — it's the ability to recreate and sustain an experience that resonates with an audience.

How does your work serve the audience?

ElevenLabs enables unique character voices to be generated in minutes. In China, the ability to manipulate AI output at an artistically high level isn't consistent yet, though the consumer experience is generally acceptable. Artists like You Zhou Xian He and Tang Haiqing are experimenting with using AI to visualize dreams and inner states — a compelling use of AI's strength in immediate, expressive content creation.

How do you define success in creating good IP?

Taste and trust (Mark): Know your audience. Understand what you want to deliver, and build trust over time.
Rights-backed experiences (Kabir): Fortnite's integration of Darth Vader demonstrates what's possible when IP is properly licensed — characters that genuinely communicate with users, creating unique interactive experiences. The value of a "clone" of an IP is high when done with integrity.
Emotional resonance with society (Song): Great IP taps into a cultural moment. Lighting producers who leverage social media with the right story and platform can build lasting emotional connections.

How is AI changing filmmaking?

Investment perspective (Kabir): A unique idea matters less than a great team. Investors look for distribution strategy, quality, and consistency — and the ability to stay current.
Problem-solving mindset (Mark): As content saturates the market, the challenge becomes removing noise. AI enables rapid iteration and audience testing that wasn't possible before.
AI as early-stage demo tool (Song): Filmmakers can now create rough AI-generated demos to pitch investors, lowering the barrier to entry significantly. The social impact of democratizing production tools is enormous — the technology should be embraced, not feared.

What matters most when seeking AI film funding?

Deliver on your promise (Mark): What you commit to your audience must be honored consistently. The craft skill — the human element — remains the key differentiator. AI shifts the risk calculus for studios, enabling faster scaling and iteration while reducing traditional production risks.
Vision and compassion (Song): Investors should look for creative leaders who have a genuine vision and deep connection to their subject matter.

In 5 years, where does AI film go?

Kabir: Technology will be widely accepted, including for voice performance work.
Song: Throughout literary history, many writers wrote without profit as the primary motivation. AI may enable a renaissance of that spirit in film — great stories told without the financial gatekeeping that blocks them today. AI makes filmmaking more equitable.

What is the relationship between IP and AI going forward?

Mark: The quality of AI-generated output determines whether IP engagement is worthwhile. If IP becomes too easily replicated, its value degenerates — protecting that value is paramount.
Kabir: The priority is preventing IP misuse. Getting the guardrails right matters as much as unlocking the creative potential.

Panel: From Canvas to Screen — The Next-Gen AI Production Pipeline

Panelists: Anthony Garcia, Qiqi Zhang, Michael Evans, Harry Zheng

What surprised you most about how people use your tools?

Qiqi: Hollywood directors think in emotional timing. They work from 10-second intention clips and care deeply about story beats — they're looking for an editing language that honors that instinct.
Michael: 90% of users have no filmmaking background. The floor has dropped dramatically, but experienced filmmakers also move much faster. Long-form consistency remains the main unsolved challenge.
Harry: Two patterns emerge — Silicon Valley users want a single end-to-end pipeline with high controllability, while film production houses need customized, proprietary agents that preserve their "secret sauce" for quality and tone.

What does an agent-based film workflow actually look like?

(Harry) The idea is that you never need to manually touch the pipeline. Agents are workers executing within a defined workflow:

Input your script → agent generates a story structure
An AI director reviews reference images and quality checkpoints
Humans adjust and approve at key stages

Production ratio is roughly 1:1 real time to generation time for most content. Commercial ads can take longer.

How do you maintain visual consistency?

Qiqi: Script → storyboard → shot list. A 5–10 minute film may require ~180 shots. Everything downstream is cutting and editing from that structure.
Michael: Extract location references per scene. Reinforce the environment look across shots, vary angles and distances, and always include close facial reference shots. Starting from the back of a character and having them turn often breaks identity — front-facing staging from the start helps. Upfront planning is essential: let the agent understand your intention before generation begins.
Harry: Three layers of consistency to solve: (1) image-level character consistency, (2) within-clip consistency — Stable Diffusion 2.0 resolves ~80% of this for 15-second clips, and (3) cross-clip consistency over an entire film — agents that actively track character state between clips are the solution.

Time savings: now vs. near future?

Task	Current	Near Future
20-min film generation	~4 min	Faster with better planning tools
Full shot-by-shot storyboard	Hours	Near-instant by end of year
Commercial content	~5 hours	—
TikTok-style short drama	~10 min	—

Open vs. closed source models?

Closed-source models currently move faster and produce better results overall. Open-source has a role, but the frontier is moving in closed labs right now.

Chroma Awards

An emerging awards platform for AI-created content across film, music, game, and other media — positioned to drive submission culture and legitimize the category.

Panel: Will AI Kill Film & TV? What's Next?

Panelists: Bryan Liu, Minh Do, Andrew Woodruff, Hardeep Gambhir

How does the AI video space feel right now?

Bryan: It's becoming invisible. Production-quality AI video is reaching the point where audiences can't tell.
Hardeep: The ceiling is higher than ever — comparable to how VFX transformed what was possible in 1992. But the floor is also lower: the worry is a world of low-quality "slop."

Where does Hollywood's center of gravity shift?

Hardeep: The bar is rising alongside the tools. High quality and low quality will both proliferate — the differentiator is always been taste.
Andrew: Legacy studios have little incentive to take risks when sequels print money. The inertia is structural.
Bryan: With 8-second attention spans, the core need hasn't changed: who can tell a good story?
Minh: The cost of AI generation will continue to fall, opening more creative space to more people.

Where is the money going?

Individual filmmakers will increasingly have the production capacity to generate real results. The expectation should be that meaningful AI-native films emerge soon — not from studios, but from individuals with vision and the tools to execute it.

Why aren't we there yet?

We don't have the right platform for great AI content to surface and find its audience.
Hardeep: We're very close to a tipping point where a single AI-generated work changes public perception.
Andrew: The "greedy algorithm" mentality — optimizing for clicks over craft — is culturally corrosive. Counter-programming matters. (Personal note: I want to make an animated short exploring the Taiwan-China neighborhood dynamic — this is the kind of story that could.)
Bryan: Audiences want to remix the content they consume. Participation and personalization are the new engagement model.

Is craft dead? What's the new norm?

Hardeep: Craft is fundamentally human creativity. Breaking Bad is a canonical example — the showrunners plotted the final season's beginning and end first, then figured out the middle. AI generates sequentially and doesn't naturally work this way. That non-linear creative thinking is still a human edge.
Bryan: Craft isn't the key variable — the audience's experience of the output is. What they feel is what matters.

GLEAM Media Studios: Creating in the Age of Algorithm

Budget pressures accelerated AI adoption for GLEAM. The key insight: if you keep doing the work with discipline and commit to the business, AI becomes an amplifier rather than a replacement. Their demo focused on using AI to generate emotionally expressive character animation — making the internal state of a character legible and compelling on screen.

Panel: The Last Mile — Global Distribution

Why is distribution so hard for AI film right now?

In the traditional studio world, the entire distribution stack is pre-built and integrated. For independent AI filmmakers, finishing the project is only half the problem — getting from "done" to "sold" has no clear path.

How do you make AI content discoverable?

Quality is the prerequisite. YouTube's early days showed that great content rises — but only if it's genuinely great. The field is small now, which is an advantage.
Format flexibility matters. During COVID, live-action stalled while animation thrived. Full-length features can follow similar pivots. Consider serializing, adapting to graphic novel format, or building for vertical (mobile-first) viewing.
Let AI help find the right channel. Match content type to distribution surface intelligently rather than treating all platforms as equivalent.

How is content distribution targeting changing?

AI-generated vs. human-made will soon be an invisible distinction.
Distribution companies that understand fandom will have an edge — harness existing fan communities.
Audiences love the creative process. Give them access: publicists, behind-the-scenes, creator dialogue.
Build a flywheel: encourage fan-made spin-offs of your IP, pay for the best ones, and let the IP expand organically.

Why does a "clean title" matter in the age of AI-generated data?

Ownership of digital clones raises serious open questions: Do digital copies of a person have legal standing? Who owns your clone if it evolves beyond your original performance? These questions sit at the intersection of IP law, ethics, and (genuinely) theology.

Talk: Hack Reality with Art

Speaker: Ben Chimney

A provocation on authorship in AI-generated creative work.

A framework for creative integrity:

Begin with intent:

What do I actually want to say?
Why does this matter to me?
Do real research — don't copy-paste AI summaries back at the world.

Reject binary thinking:
The most interesting work lives outside yes/no, true/false. Bring your own experience, judgment, and contradiction. Connect your own dots.

Core insight: What is yours? When your real experience, perspective, and taste are embedded in the work — that's what makes it yours. The tool is irrelevant.

Panel: The Attention Economy — AI in Commercials

Panelists: Auriel Wright, Daishu, Amy Zhang, Phillip Meier

How does AI compare to traditional commercial production?

Daishu: AI transforms the iteration cycle — it shifts resources toward what you actually care about rather than what production logistics demand.
Phillip: AI is a layering tool. Start rough, refine with each pass. The visualization capability makes the creative exploration more accessible.
Amy: Traditional commercial production starts at $100K+ before crew. AI fundamentally changes the risk profile — and the IP safety conversation changes with it.

How does AI help short-form drama specifically?

Story remains the critical variable. High-frequency short-form content enables rapid data feedback loops and fast ROI, which makes it easier to fund genre and fantasy content that would otherwise be too risky.
Low-quality content works in the attention economy as a psychological mechanism — but the deeper game is getting to people's core emotional needs, not just their reflexive attention.

Personalization and the attention framework

The asset's look — visual identity, character design — can be tuned to viewer profile. Imagine IP-branded phone cases featuring a character variant designed around personal taste. That's the direction.
The "3 needs" framework for content that matters: guilty pleasure, artistic requirement, and what actually matters to a person. Most content serves only the first. The most powerful content reaches the third.

What drags attention?

Conflict — compressed and intensified in short-form content. Short drama is condensed storytelling under pressure.

What makes AI-generated content genuinely good?

Control — the tool understands your intention even from a vague input.
AI as conversation partner — the best results come from dialogue with the tool, not commands. Something unexpected always emerges from the exchange.