The Next Rembrandt - ‘the’ Pioneering Art Meets AI Statement, Anno 2016
Going back into the archives, this innovative campaign was ahead of its time - and remains a wonder… An AI-constructed and 3D-printed painting, this new ‘Rembrandt’ made history - causing a media storm and leading to a cultural shift in perceptions of AI and tech merging with art. The Next Rembrandt is an iconoclastic take on the works of the Dutch Old Master, and took the possibilities of AI to the next level. The award-winning campaign opened up a conversation about the power of AI that still resonates today - almost eight years later.
Through ‘The Next Rembrandt’, we kicked off a global discussion about how technology will intersect with creativity in the years to come and pushed the boundaries of AI’s application to fine art. This art meets tech creative revisited the archive of one of the Netherlands, and the world’s, most venerated artists, and the application of AI to research into Rembrandt’s technique was met with applause and criticism in equal measure. As innovative as it was controversial, the piece not only became a huge news story but also raised important questions about authenticity, art, and technology, the real value of art, and the role of AI in creativity.
In 2016 I worked on one of my all-time favourite campaigns ever, one of the most innovative and slightly mind blowing projects I’ve launched - and one that caused a culture shift in how people thought about creativity and the advancement of AI and machine learning.
I worked closely with creative chief and innovation legend, Bas Korsten, and his agency J. Walter Thompson (now VML) to create the PR strategy and launch campaign for ‘The Next Rembrandt’ – a new portrait comprising 148 million pixels, created with machine learning based on 168,263 fragments from 346 paintings attributed to Rembrandt.
Watch the launch video that explains how it was created, and the PR campaign around it here
Commissioned by Dutch bank ING, the brief from the client was “to match ING’s innovation in the banking world in the domain of art and culture. This brief from such a high-profile client set the bar high, but we raised to the challenge and conceived the very first AI-generated artwork that was totally new, but based solely on original data sources.
Such a radical proposition attracted the cooperation of leading experts including data scientists, developers, engineers, and art historians from Microsoft, Delft University of Technology, the Mauritshuis in The Hague, and the Rembrandt House Museum in Amsterdam.
The appeal was clear: What if the technology of the future could bring this Old Master back from the past? Could AI, machine learning and tech distil the essence of Rembrandt into data? Turning strokes of genius into lines of zeros and ones?
Authenticity vs AI?
‘Is this really a Rembrandt?’ This question has historically caused a media furore due to the various forgeries that have hung in museums for decades, fooling art admirers and historians alike. So imagine the impact when perhaps the most convincing forgery yet appeared, and it was created without a single human hand or brushstroke being applied to the canvas. This was the supreme deep-fake, but it was made in 2016 before the term entered our vocabulary. Even for those claiming to be able to tell the difference, the answer to the question was still not clear. It was hard to argue that it wasn’t a Rembrandt – after all, it was built from data gathered from the Old Master’s entire oeuvre.
How was it not a Rembrandt, if made solely from Rembrandt’s body of work?
Part of that appeal was that of re-examining the archive of one of the most revered painters in the world and transporting his art and technique into the present day through AI. Though his global fame is undisputed, many forget that Rembrandt was a true innovator, using light and shadow in new and truly groundbreaking ways, bringing to portraiture an intimate, poetic depth not seen before. His work went on to profoundly influence other artists, including Vincent van Gogh, who was famously transfixed in 1885 by The Jewish Bride. These days we call Rembrandt an Old Master, but at the time he was the cutting edge.
The project therefore also represented a unique challenge for art historians. What exactly was it that made Rembrandt’s paintings so special? While always likely to draw the ire of Rembrandt puritans, The Next Rembrandt was also a form of devoted fan service – distilling, with numerical precision, exactly what makes a Rembrandt a Rembrandt.
The unique technological challenges were overcome by an army of art academics and AI experts through extensive studies of Rembrandt’s use of geometry, composition, and painting materials.
__________________________
168,263 painting fragments were collected, and then a deep neural network algorithm was trained to upscale and standardise each scan to equally high resolution. To mimic the painter’s uniquely human touch, the data was analysed to recreate the quintessential Rembrandt portrait subject. After classifying over 400 faces, the data concluded that the subject should be around 30 to 40 years old, male, wearing black clothing, a hat, and facing the right. Over 6,000 facial landmarks were then used to classify ‘typical features’, and these were then assembled into a fully formed face and bust that respected Rembrandt’s individual use of proportion.
Producing the object was its own work of art because, in order to move from a 2D image to a 3D printed painting, the team had to add another dimension with three layers: canvas, ground layer and brushstrokes. X-ray scans and digital cross-referencing those thousands of fragments recreated that distinct Rembrandt texture. Finally, thirteen layers of a special paint-based UV ink were used with Canon’s high-end, custom-made 3D printer – the first time ever new work was printed in such a way.
The reaction from the public and the media was immediate: after an exclusive in The Guardian, The Next Rembrandt was the top news item of that week, creating headlines across the whole world and popping up on prime-time news shows. Blurring the boundaries between technology, art and humanity, the ‘painting’ was immediately divisive.
After all, it raised so many fundamental questions.
It was with these questions and themes that we amplified the whole project. Writing the PR strategy, producing all the storytelling and coming up with the launch concept, our communication direction made this one of the art stories of the year.
Was this art? This was no human expression. And while by no means the first forgery, The Next Rembrandt was nevertheless a pioneering case of technology truly trampling on the sacrosanct, creating a genuine culture shift by highlighting the insightful power of technology several years ahead of AI’s current ubiquity.
Could artists be replaced by machines? Here, AI was stepping on man’s most precious domain – creativity. This hit a nerve across the art world and beyond.
What was it worth? It was not technically ‘an original’. But if you converted the value of The Next Rembrandt into such modern-day metrics as media value and social media impressions, the piece was priceless.
And of course Was this a Rembrandt? Ultimately, The Next Rembrandt was of him, but not by him. Every brush stroke his, but not by his hand.
The results
1400+ articles written
Globally trending on Twitter
1.8 billion media impressions
Increased stock price for ING & Microsoft
Cannes
Cannes Lions - Cyber: Grand Prix, Silver Lion, 4 Shortlists
Cannes Lions - Creative Data: Grand Prix, 2 Silver Lions, Bronze Lion, Shortlist
Cannes Lions - Innovation: Innovation Lion
Cannes Lions - Outdoor: 2 Gold Lions
Cannes Lions - Digital Craft: Gold Lion, 2 Bronze Lions, Shortlist
Cannes Lions - Direct: Silver Lions, 2 Shortlists
Cannes Lions - Promo & Activation: Silver Lion
Cannes Lions - Design: Silver Lion
Cannes Lions - Entertainment: Bronze Lion, Sh