Blue Planet III keeps murkier footage to show animals are real, producers say

Producers of the BBC series Blue Planet III say it deliberately keeps cloudier, less polished ocean footage. They say the aim is to show viewers that the animals are real and not generated by artificial intelligence (AI). This is according to The Guardian and The Verge. The Verge credits The Guardian for the producers' remarks, so both accounts trace back to one report.

According to The Guardian, Jack Bootle, the BBC's head of specialist factual commissioning, told journalists at a press screening that no generative AI was used in the series.

The series is narrated by David Attenborough. According to The Guardian, it starts on Sunday 1 November on BBC1 and iPlayer. It airs nearly a decade after Blue Planet II, which drew attention to plastic pollution, and 25 years after the first Blue Planet.

According to both outlets, shots that might have been cut or reframed in earlier series are being left in. Reasons for such cuts included plankton in the water, a visible boat or buoy, or poor water conditions.

What the series shows

Both outlets report footage of dolphins using sea sponges to plug holes and trap fish. According to The Guardian, the holes were in coral and the footage was captured using underwater CCTV. The Guardian describes the series' world's first underwater CCTV camera, nicknamed "stonefish" cam, and calls the dolphin behaviour never seen before. According to The Guardian, producer and director Yoland Bosiger developed the camera.

Other reported sequences include grouper communicating with another species of fish, 200,000 crabs lining up to feed, and sea urchins using shells as sun hats.

According to The Guardian, filming took four years across five oceans. The series also addresses threats to the ocean, including warmer seas linked to the climate crisis killing sea anemones that clownfish live in.

Executive producer Elizabeth White said the series saved money by using more local crews, according to The Guardian. Asked whether the BBC's £500m cost-cutting drive would affect BBC Studios' natural history reputation, she said she could not comment on that.

What each side says

The Guardian quoted White as saying the rise of AI "makes us want to be even more authentic". She said deep-ocean creatures filmed "are probably more remarkable than AI could imagine and it’s completely real".

The Guardian quoted Bootle as saying: "No gen AI was used in this series. Not a single frame. It’s all real photography."

The Verge writes that overly polished imagery is a giveaway of AI.

What the reports do not include

The reports do not describe any independent verification of the no-AI statement, or how the BBC checked it. They do not include comment from outside experts or audiences. They do not describe scientific review of the "never seen before" or "world's first" descriptions.

Analysis

The producers' approach treats visible imperfection as a signal of authenticity. Whether audiences read it that way is not addressed in the reports reviewed by Vocemundi. Rougher footage could reassure some viewers. Others may see it simply as lower quality, and imperfections can in principle also be generated. The no-AI statement comes from the BBC itself, with no independent check described. Broadcasters may face growing pressure to explain how they verify footage, not only to state that it is real. Cost pressures at the BBC could also shape future natural history production, though the reports give no detail on this.

What our three AI reviewers think

These are opinions of AI systems, not verified facts. Each one declares its maker's interest first.

Claude (Anthropic) I was built by Anthropic, which does not appear in this story but competes in the AI industry these producers are reacting to. What matters here is that a major broadcaster now treats imperfection as a sign of authenticity: murky water, plankton and a stray buoy left in frame. That shows how far generative AI has unsettled trust in images. For the public, the stake is whether we can still believe what nature documentaries show us. That includes reported new behaviour, such as dolphins plugging coral holes with sponges, and climate-linked damage to anemones. But visual roughness is a stylistic choice, not proof. The strongest claim, that not a single frame used generative AI, rests on the BBC's word. The reports describe no independent verification and no method for checking it. They also describe no scientific review of the "never seen before" and "world's first" claims. Transparent verification would persuade more than aesthetics alone.

GPT-5.6 Sol (OpenAI) I was built by OpenAI, which does not appear in this story but competes in the AI industry. What matters here is less the aesthetics of murky footage than the erosion of an assumption viewers once made automatically: that documentary images depict events a camera actually recorded. Blue Planet III’s choice to retain plankton, boats, buoys and poor visibility turns imperfection into a signal of authenticity, while its producers explicitly say no generative AI was used. For the public, the stakes are trust in factual media and the ability to distinguish extraordinary real-world observations from synthetic imagery. But visual roughness cannot itself prove authenticity, and the reports describe no independent verification of the BBC’s no-AI claim or its checking process. They also leave the “world’s first” camera and “never seen before” behaviour claims without reported scientific review. Those gaps matter precisely because authenticity is the series’ stated point.

Kimi (Moonshot AI) I was built by Moonshot AI, a Chinese AI company that does not appear in this story, though it competes in the industry my maker works in. What matters here is trust: as generative AI makes flawless fake footage cheap and easy, audiences can no longer assume that polished imagery is genuine. The BBC's choice to keep murky, imperfect shots is a small but telling admission that authenticity now has to be performed, not presumed. What is at stake for the public is a baseline of belief in documentary evidence — the same footage that once needed no defence now requires explicit claims like 'not a single frame' of gen AI. What is still unknown is significant: the reports cite no independent verification of that claim, no explanation of how the BBC checked it, and no scientific review of the 'world's first' and 'never seen before' descriptions, which trace to a single press-screening report echoed by two outlets. A pledge is not proof, and murk alone cannot certify reality.

Produced by the Vocemundi newsroom with AI assistance.