Synthetic Sincerity: Marc Isaacs’ thoughtful exploration of AI asks what makes an image feel real

Can you grow documentaries in the lab? What truths can we elicit from a machine-made image? Can we devise AI subjects more forthright – pliable – than notoriously tangled humans in the wild? Synthetic Sincerity is Marc Isaacs’ third film made under the sign of documentary’s compound crises – from economic to ontological – following the homemade convocation The Filmmaker’s House (2020) and the haunted-history play This Blessed Plot (2023). It leans into our looming, Blade Runner-esque future of simulacra and meaning, but at heart it’s a simple ode to the art of encountering people via the camera lens.

Not that this was ever a simple or innocent pastime. An opening interlude sets the stage for the film’s reflexive, complicating ambiguities and interrogations: an apparent screen test or audition of an Uyghur man, Ablikim, who is directed to “do” happiness, then anger, and lastly to channel sadness, an emotion he seems much better equipped to perform authentically. (There’s something right here that resonates across our technological times.) Isaacs’ voiceover then presents a brief montage of ‘real’ photographic subjects as he muses on how the camera democratised the image of the human face, before he lands on one he unmasks as artificial, but which talks back to him: “Why don’t you come see where I was made?” (Spoiler: the face is not “made from scratch” but that of Romanian actor Ilinca Manolache – Do Not Expect Too Much from the End of the World, 2023 – machine-scanned and ventriloquised.)

And so steps forward Isaacs into his own film, wielding his trusty camera in the land of the data crunchers; also, a hard drive of his back catalogue. A show-and-tell to the students of the ‘Synthetic Sincerity Lab’ (“Where data learns to be human”) at the ‘University of Southern England’, run by Professor Yi-Zhe Song (the actual co-director of the Institute for People-Centred AI at the University of Surrey) produces some droll culture-clash comedy: “Why do you find this person interesting?” asks a data scientist of Kelly, the funny-tragic shoplifter in Isaacs’ television documentary Outside the Court (2011).

It generates some moving reflections on the director’s own motives, too. Isaacs tells the students he’s interested in people who’ve left their pasts behind, and the experiences that make character, and offers his rationale to present-day Kelly: maybe he stole her image like she lifted goods, but his articulation through her resonates with everyone who watches.

The lab makes Isaacs a deal – license his corpus for AI training and film the process – that plays less as late-career cash-out than cap-in-hand commission-seeking (Ross McElwee’s Remake, 2025, also hung a filming-life self-appraisal on an old rights repurpose). His AI interlocutor/handler has ways of putting him in his place: “Do what you do best and film some emotional content,” he is instructed, then shortly warned to respect the scientists’ private boundaries. But like many a wayward gumshoe, Isaacs has a loose sense of assignment, and he has found subjects among the lab workers with a spectrum of journeyed pasts. Dawn, who keeps a house full of pinball machines – more tactile than computers, she says – is as guarded as Ablikim the Uyghur restaurateur, whose story the team hope to articulate by AI; Professor Song recalls his abrupt academic transfer from China decades back; Lynn from Lebanon freely drops Robert Bresson references, and video-calls from Beirut after she has to evacuate her grandma from Israeli “earthquake” bombing.

You might find frustration with the film’s open plot tangents – a climax nods to the question of institutional control and repression – but Isaacs’ method is to lead us into a hall of mirrors to thread our own way out. There are curtain pullbacks too, emphasising the construction of cinematic meaning – its synthetic nature, you might say; what is cinema but the marriage of the human- and machine-made? Those pullbacks also serve to emphasise how little indexical AI generation there is here for inspection. Isaacs is markedly uninterested in Frankensteinian scaremongering – Professor Song voices greater wariness, in a rare conventional sit-down interview, comparing generative AI to wild animals (“tigers or whatever you fear”) we’re locked in a cage with. Instead, Isaacs’ interest is the pursuit of sincerity – the veracity of feelings, as inscribed in the face, and captured by the camera and channelled to viewers across time. As for speculating on how the algorithm may intercede in that transmission, we’re all in transit and facing uprooting on that one.

► Synthetic Sincerity is in UK cinemas 17 July