One of the most useful lessons about AI in entertainment may come from a filmmaker deciding not to use it. Source: Venice Film Festival listing for Flesh Impact
At the Venice Film Festival, Maggie Gyllenhaal described experimenting with AI while making the short film Flesh Impact.
The production had licensed Marilyn Monroe material and tested AI to transform Dakota Johnson’s performance into Monroe.
Gyllenhaal ultimately rejected the AI-generated result.
Her reason matters.
The technology removed something from the performance that the filmmaker considered more valuable than the technical transformation: its humanity.
Responsible AI does not mean maximum AI
The entertainment industry is currently surrounded by claims about what generative AI could make faster or cheaper.
Some of those claims will prove correct.
AI can already support development, visualisation, localisation, workflow automation and other production tasks.
But capability alone is not a creative justification.
The correct question is not:
Can AI do this?
It is:
Does AI make this particular piece of work better?
Those are very different standards.
Human judgement remains the final gate
Gyllenhaal’s experiment is useful because the technology was actually tested rather than rejected in principle.
The filmmaker then made a judgement about the result.
That is a credible model for creative AI adoption.
Experiment.
Evaluate.
Compare.
Retain what improves the work.
Reject what diminishes it.
Human creative authority remains the final gate.
Rights clearance does not guarantee creative success
There is another important point.
The production reportedly used licensable AI and licensed Monroe material.
That addresses part of the rights question.
But a rights-cleared AI workflow can still fail creatively.
Governance and creative quality are therefore separate tests.
A responsible entertainment workflow needs both.
Is the use authorised?
And is the result good enough?
AI should strengthen the creative process
The strongest future for AI in entertainment is unlikely to come from using the maximum possible amount of AI.
It will come from using the technology selectively where it creates genuine value.
That can mean faster development.
Better previsualisation.
New production economics.
More accessible experimentation.
Or entirely new forms of storytelling.
But the purpose of the technology should remain the work itself.
At SOS Entertainment, our position is straightforward: technology should expand what creators can build without confusing automation with creativity.
The future of entertainment may use significantly more AI.
The best entertainment will still require somebody with the judgement to say no to it.
Apply this intelligence to an accountable commercial decision.
Discuss human-led production with SOS