AD FIELD NOTESAI advertising · the work behind the output

Case files / machine learning cultural campaign

ING The Next Rembrandt: use the generated object as the media event

ING and partners used data analysis, deep learning and 3D printing to create a new Rembrandt-style portrait, then took the object into public outdoor display.

Campaign period: April 2016Source published: 11 May 2016Archive entry prepared: 15 September 2026

Retrospective coverage, prepared in September 2026 from the original record.

Outdoor ING display presents The Next Rembrandt portrait in Amsterdam.
The generated portrait shown in ING's outdoor touring display. ING / J. Walter Thompson Amsterdam / Microsoft; image published by JCDecaux Netherlands.

Reported execution: a painting became the campaign artifact

ING's 11 May 2016 announcement says The Next Rembrandt was unveiled on 5 April and then toured Dutch cities in illuminated outdoor displays. It describes deep-learning algorithms and facial-recognition techniques applied to 346 known Rembrandt paintings, followed by a 3D-printed portrait. ING names JWT Amsterdam, Microsoft, university and museum partners alongside the bank.

The public work was not just a visualization of a technical process. The portrait, and then its staging in streets, gave people a physical object around which to have the authorship conversation. That is the advertising execution this archive records.

Our reading: make the output carry the proposition

The most persuasive choice was to let the result be inspectable. A claim that a model had learned an artist's style would be abstract; a framed portrait in an outdoor cabinet makes the claim concrete enough to admire, doubt or discuss.

This is a useful pattern for technical brands: build an artifact that embodies the proposition, then give it a setting where ordinary attention can reach it. The artifact must still earn that attention. A complicated process alone does not make a relevant brand idea.

Credit the system without flattening the contributors

The ING account explicitly names data scientists, developers and Rembrandt experts. That matters because the public shorthand ‘AI made a Rembrandt’ obscures choices about the corpus, subject, style constraints, printing and exhibition. Those choices are the work that made the result legible.

For a related campaign, a compact methods note can improve the story rather than weaken it: what inputs were licensed or permitted, what the model was asked to do, and what experts decided afterward. The goal is not to turn an ad into a lab paper; it is to avoid a magical claim of autonomous authorship.

Evidence boundary: cultural attention is not product proof

The primary announcements establish the dates, partners, technology description and outdoor rollout. They do not provide a public controlled measure connecting the work to brand preference or business outcomes.

Treat The Next Rembrandt as an unusually clear example of AI-led cultural communication, not proof that a generated cultural object will deliver the same value for every sponsor.

Sources & limits

ING and JCDecaux announcements document the project and touring display; they are not independent evidence of brand effect.

  1. ING Newsroom — The Next Rembrandt on tour — 11 May 2016.
  2. JCDecaux Netherlands — ING exhibits The Next Rembrandt — 9 May 2016.

Send a correction with the passage and supporting source.

The weekly field note

One useful note for the next brief.

Campaigns worth studying, the work behind them, and questions to bring into your next creative review. Delivered weekly.

How we handle your email