Maggie Gyllenhaal’s latest directorial effort reportedly included an attempt to digitally recreate Marilyn Monroe using artificial intelligence, but the result has generated far less buzz than Dakota Johnson’s recent take on the Hollywood icon. The project, which has been the subject of quiet industry chatter since Gyllenhaal first floated the idea, has now become a talking point for how studios and filmmakers are approaching the legacy of classic stars in the age of generative technology.

According to insiders, Gyllenhaal was approached with the suggestion to use deepfake-style AI to bring Monroe back to the screen for a specific sequence, presumably to evoke the actress’s signature presence without casting a lookalike. The request appears to have been part of a broader push to leverage AI tools for period-specific authenticity, a trend that has been slowly creeping into major productions over the past year. However, early test footage and behind-the-scenes reactions suggest that the Marilyn recreation fell flat, lacking the subtle human qualities that made Monroe so magnetic in the first place.

The comparison to Johnson is almost inevitable. Johnson’s own recent project, which touched on Monroe’s life or aesthetic, has been praised for its grounded, performance-driven approach rather than relying on digital trickery. While Johnson did not impersonate Monroe in a traditional biopic sense, her portrayals have captured a certain vulnerable glamour that critics say feels more alive than any algorithm-generated approximation. The contrast has become a shorthand in Hollywood circles: one actress brought Monroe back through craft, the other through code - and the audience can tell the difference.

No official footage or stills from Gyllenhaal’s project have been released, and the film itself has not yet been formally matched to any release date or title, leaving much of this story in the realm of production gossip. Still, multiple sources with knowledge of the shoot have indicated that the AI-generated Marilyn was intended for a dreamlike or memory-based scene, where a real actress might have disrupted the surreal tone the director was aiming for. That rationale, while conceptually interesting, did not survive contact with the final image.

“It looked like Marilyn, but it didn’t feel like her,” one crew member reportedly said, speaking on condition of anonymity. Another source compared the effect to a wax museum figure that had been animated by a novice - technically convincing in shape, but devoid of the spark that makes a performance memorable. The uncanny valley, it seems, was not crossed, and the sequence was quietly revised. Whether the scene was cut entirely or re-shot with a human performer remains unclear, but the backlash inside the production team was loud enough to force a change of direction.

This is not the first time AI has been used to resurrect deceased actors, but Marilyn Monroe presents a unique challenge. She is one of the most photographed and filmed women of the 20th century, which means there is ample data for an AI model to learn from. That very abundance works against the technology, too. Audiences have deeply ingrained memories of Monroe’s specific gestures, her vocal flutter, the way she held a gaze. When a digital recreation gets even a detail slightly wrong, the failure is glaring. With a living actress, there is room for interpretation; with a digital avatar, the expectation is exact replication, and anything less reads as a forgery.

Johnson’s advantage, according to film critics, lies in her ability to channel the idea of Monroe without attempting a carbon copy. Her performance leans into the tragedy and wit that defined Monroe’s public persona while remaining unmistakably Johnson on screen. That approach allows for a creative dialogue with history rather than a sterile duplication of it. The comparison between the two methods - organic interpretation versus synthetic reconstruction - has quickly become a case study for how Hollywood will handle its backlog of beloved stars going forward.

Gyllenhaal, who has built a reputation as a thoughtful director with a keen eye for female psychology, has not publicly commented on the AI experiment. Industry observers note that she is unlikely to address the rumor directly, as doing so would acknowledge a misstep. However, the fact that the story has leaked suggests that the production team itself may be divided on the ethics and efficacy of the approach. Some younger crew members are reportedly enthusiastic about the future possibilities, while veterans remain skeptical of any tool that replaces a living performer.

The broader conversation extends beyond this single project. Studios are already exploring AI licensing deals with the estates of deceased celebrities, aiming to secure image rights for generations to come. Monroe’s estate, which has historically been protective of her likeness, has not issued any public statement regarding the reported Gyllenhaal collaboration. If the sequence was ultimately scrapped, it may be because the estate itself refused to approve the final render, another possible point of friction that has not been confirmed.

For now, the takeaway in Hollywood is clear enough: Dakota Johnson’s approach felt human; Gyllenhaal’s AI recreation felt synthetic. As the industry continues to debate whether digital resurrection is a creative tool or a creative crutch, this incident serves as a cautionary tale. Technology can mimic a face, but it cannot yet replicate the vulnerability, timing, and instinct that define a star’s unique energy. Maggie Gyllenhaal may have learned that lesson the hard way, but her willingness to experiment ensures she remains a director to watch - even when the experiment fails.

Source: Hollywood Reporter