Summary: Jon M. Chu says one of Wicked’s best moments—the spontaneous wink from Cynthia Erivo—could not have been manufactured by artificial intelligence. He frames that claim with three threads: the creative value of improvisation on set, the real-world lessons he learned early in his career about direct fan connection, and a cautious but practical view of where AI helps and where it cannot replace human discovery. This post breaks those threads down, shows what filmmakers and marketers should copy from Chu, and offers concrete steps to keep room for the unplanned while taking advantage of AI’s strengths.
Why a spontaneous wink matters
A wink is only a wink if it arrives in the moment. A scripted wink reads like a line. A spontaneous wink lands like a revelation. Chu says that when Cynthia Erivo winked while putting on her witch cape, the shot became an icon because it was unplanned. Repeat that: unplanned. Unplanned moments carry a kind of truth actors and audiences sense immediately. They signal presence, risk, and a human mind reacting in real time.
Think about your own reaction when a performer does something you didn’t expect. Did you lean in? Did you smile? That is the value we’re measuring. It is not an abstract aesthetic preference. It is a measurable engagement spike: attention, shared clips, conversation. That’s why the wink matters beyond charm—it influences how an audience connects with a character and with the film.
Improvisation versus preprogrammed perfection
AI can simulate patterns, generate alternatives, and polish phrasing. It cannot, however, generate the tacit know-how that lives in a room under pressure. Tacit know-how includes timing, breath control, eye contact, and the subtle negotiation between actor and camera. You can ask: what would it take for a machine to mimic that tacit layer? The right answer is not just more data—it is bodily presence.
If every gesture, camera move, and line is locked before cameras roll, discovery disappears. You lose the moments where collaborators find solutions together. You lose the capacity to change a performance based on an actor’s intuition. You also lose the small human errors that reveal character. That is the downside of demanding perfection in advance.
The Bieber lesson: fan connection starts before filming
Chu traces a clear line from shooting Justin Bieber’s Never Say Never to understanding how to build a story that lives before and after the film. When Bieber posted a joke that Chu was “following him around,” tens of thousands of followers showed up for the director overnight. What does that demonstrate? Direct connection scales fast. The lesson: narrative begins in the real world, with small acts of access and trust.
Fans don’t only watch a movie; they inhabit the social threads around it. For Wicked, the cast’s off-screen chemistry became part of the product. That organic bond turned marketing into proof rather than promise. In short: real relationships produce credible publicity.
Silicon Valley helped Chu, but not by replacing art
Chu credits tech-savvy customers at his parents’ restaurant for giving him early tools—computers, video cards, software—that helped him get into USC. That generosity mattered. It gave him technical advantage. But he never suggested tech replaces craft. He owes his start to tech access, and he feels responsible to the region. That’s a model of reciprocity: the community helped him, and he now gives back through work that lifts profiles and pays wages.
This is a nuanced stance. Tech helped level the playing field for one creative kid, but tech alone does not create the on-set chemistry that produces magic. The social infrastructure—trust, mentorship, real human exchange—remains central.
Where AI is useful, and where it is not
Chu is open to AI for gathering information, organizing research, and even generating starting points. Use AI as a bench player: efficient, tireless, precise. But he draws a line at relying on AI for the deciding call on-cam. Why? Because the call depends on embodied judgment: timing, feeling, the micro-adjustment that an actor makes to a line once the lens is on them.
Ask yourself: which parts of my process benefit from automation, and which parts require people in the room? Let that question shape your workflow. When you stop asking that question, you outsource judgment you cannot get back.
Practical rules to protect improvisation on set
Apply rules that preserve discovery but keep production efficient. Here are concrete steps Chu’s view implies:
– Reserve a few takes for free play. Let the director call a “play” take where actors can try unscripted beats.
– Build slack into the schedule for surprises. A 15–30 minute cushion pays for one magic moment that lifts the whole film.
– Use AI to test options off-set—line variations, blocking ideas, mood boards—so the set is a place to try rather than plan every move.
– Encourage “No” as a tool. Let crew and cast say no to over-controlled plans. Saying no defends space for discovery.
Directing as negotiation: Chris Voss tactics for set leadership
Directing is negotiation. Use calibrated, open questions: What do you think would make this line feel truer? How can we make that beat land without losing pace? Those questions invite creative ownership. Mirror actors subtly: repeat a key word they use—”so you want urgency? Urgency.” Label emotions: “It sounds like you’re nervous about losing the moment.” That lowers defenses.
Invite silence. After you ask a question, hold your posture and wait. The pause gives the actor room to search for something real. Ask open-ended questions, mirror, label, and then be comfortable with silence. That’s where discovery often shows up.
Marketing and the long story
Chu’s view of storytelling stretches from pre-production social posts through the press tour and beyond. Fans follow people more than plot beats. The cast’s real friendships became a marketing asset because they were true. What marketing should copy from Chu is this: document authenticity, don’t manufacture it.
How can you show authenticity without exploiting it? Start by asking: what parts of the process are worth sharing, and how will sharing affect those relationships? Share the right small moments—backstage laughter, quiet rehearsal observations—not staged set pieces. Social proof matters; when fans see genuine bonds, their commitment deepens.
Balancing commercial pressure and artistic risk
Chu compares the cast bonding under pressure to developers in Silicon Valley delivering a product. That comparison is precise. High stakes compress teams and create trust. But pressure can also produce safe choices. The remedy is leadership that protects space for risk even under deadline pressure. Ask: what will we accept missing to gain a moment of truth? When you define that limit, you enforce standards without killing discovery.
You will want predictable outcomes; you will fear wasted time. Those are valid fears. Yet, refusing to allow improvisation because of fear is a different risk: you may ship a polished product that feels hollow. The question you must answer is: which risk are you willing to take?
How to test Chu’s claim in your own work
Try a small experiment: pick a scene and shoot it twice—one fully rehearsed and blocked, one where actors are invited to play for two takes with no scripted small gestures. Compare. Which footage sparks the stronger reaction from a test audience? Which clip gets shared more on social? Use those metrics to set policy.
If the spontaneous take wins, you have empirical support for keeping improvisation in your process. If the scripted take wins, examine conditions: did you give actors the psychological safety to try? Did production constraints clip possibility? The test teaches either way.
Closing reflection
Jon M. Chu’s stance is simple and clear: technology helps, but it cannot replace human discovery happening in the room. That does not mean rejecting AI. It means using it where it improves workflow and keeping human judgment where it matters most. If you’re making films, building campaigns, or shaping products, keep one corner of your process designated for the unplanned. Protect it with schedule slack, permission to say no, and leadership that waits for the moment to arrive.
Which unplanned moment would you protect on your next project? How will you design for it? What would happen if you stopped scripting every beat? Ask those questions out loud on your next call, then hold your silence and listen for the answer.
#Wicked #JonMChu #CynthiaErivo #ArianaGrande #Filmmaking #AIinFilm #Improvisation #FanConnection #MarketingStrategy
Featured Image courtesy of Unsplash and melanie nobaru (s3IdgO9j4Fs)
