Steven Soderbergh's John Lennon film arrived at Cannes with two facts that cannot be separated: it uses AI-generated imagery, and it did so with Lennon-family cooperation. The paper's Monday account of why the Cannes AI Lennon film needed a rights-holder answer treated consent as the missing question. Tuesday's fuller record sharpens it. Taste is not the only issue. Authority is.
TIME's Stephanie Zacharek reviewed John Lennon: The Last Interview as a documentary weakened by what she called "pointless AI gimmicks." The film covers the Dec. 8, 1980 radio interview Lennon and Yoko Ono gave at the Dakota, hours before Lennon was killed, and TIME reported that the credits name Meta as a technology partner. [1]
El Pais added the production architecture. Soderbergh used family album photos and AI to illustrate the radio interview, compiled more than 1,000 images, and had the Lennon family "on side." It also reported that Soderbergh sought support from Meta, a financial partner and Cannes co-sponsor, to create what he called "thematic surrealism." [2]
That makes this more than the usual AI-art quarrel. A living artist can approve a tool, reject it, sell it, or denounce it later. A dead artist cannot. Estates, heirs, archives, labels, and festivals stand in. The legitimacy of the image therefore depends on both law and trust: who owned the material, who granted access, who approved transformation, and who benefits from the revived likeness.
TIME's aesthetic objection is serious. Zacharek wrote that the film works best when Soderbergh sticks to classic documentary construction: voiceover and archival material. She argued that the AI flourishes render the film less distinctive, not more. [1] That criticism belongs in the review. It does not exhaust the cultural problem.
El Pais reported that the full interview had not previously been accessed in its entirety and that Soderbergh respected the chronology of the conversation. [2] It identified the KFRC participants, the role of family photographs, and the use of AI for abstract passages that otherwise lacked visual material. [2] Those are consent facts as much as production facts.
X will make the argument moral because X is built for moral compression. AI Lennon is desecration. AI Lennon is authorized art. AI Lennon is proof that studios will never leave the dead alone. AI Lennon is proof that estates can manage legacy. MSM will make the argument aesthetic because Cannes reviews are built for taste. The paper's job is to hold the harder middle: permission can be real and still not settle whether the use is good.
Meta's presence makes the question institutional. A tech partner at Cannes is not just a tool vendor when the tool changes what counts as archival image-making. Vanity Fair's Cannes live file noted that John Lennon: The Last Interview was already generating controversy for its use of AI-generated imagery. [3] The controversy is not a side effect. It is the new condition of festival documentary.
The answer cannot be that estates always solve the problem. Estates are not the dead. Nor can the answer be that any AI use is theft once heirs cooperate. The useful line is more demanding: disclose the source material, disclose the approval chain, disclose the technology partner, and let viewers judge both consent and craft.
Soderbergh may have made a film that is legally permitted, family-enabled, technologically current, and aesthetically weaker for the very device that made it news. That is not a contradiction. It is the AI culture question Cannes has been trying to avoid by talking only about taste.
-- ANNA WEBER, Berlin