Welcome to the Legal Palette

An art law blog

Who Painted Rembrandt? Copyright and Authorship of Two Rembrandt Portraits

Who Painted Rembrandt? Copyright and Authorship of Two Rembrandt Portraits

Rembrandt Workshop.  Portrait of Rembrandt,  (1650), oil on canvas. National Gallery of Art, Widener Collection, 1942.9.70.

Rembrandt Workshop. Portrait of Rembrandt, (1650), oil on canvas. National Gallery of Art, Widener Collection, 1942.9.70.

Who is the artist/author of a painting?  The simplicity of this question may seem facetious at first; but in the context of copyright, it triggers a complex analysis.  The foundation of copyright law outlined in the U.S. Constitution grants authors exclusive rights to their works for a limited time.  But both the Constitution and the Copyright Act fail to define “author.”  Case law has attempted to fill this gap by granting the person who has created the work or with whom the work originates the status of author.  Art similarly grapples with attribution issues as it is not always obvious who was holding the brush and palette.  With the absence of any documentary evidence, art historians rely on the composition, technique, style, and scientific analysis of a painting to establish the artist’s identity.  Two portraits with different connections to Rembrandt suggest that in order to determine the copyright author and artist today, one must look to the artistic practices and customs of almost four hundred years ago.

Rembrandt’s Portraits

Rembrandt van Rijn (1606-1669), one of the most famous artists of Holland’s “Golden Age,” was a prolific painter, draftsman, and etcher.  Rembrandt became a prominent portraitist while living in Amsterdam, with his first portrait appearing on the market in 1631.  Portraitists during the Dutch Golden Age sought to paint a more or less accurate likeness of a specific person.  (Ekkart 2008, 17).  Rembrandt’s direct observation of his subjects enabled his portraits to appear very natural in composition as well as technique.  (Ekkart 2008, 31).  Rembrandt often adjusted the color and transparency of his paints by layering a limited number of pigments to achieve the desired effect, like using lead white with high impasto for textured white ruffs.  (Broos).  With his technical concern for light, space, atmosphere, and texture, Rembrandt’s portraits serve as “psychologically penetrating descriptions of behavior.”  (Liedtke 2007, 547). 

Like many of his contemporaries, Rembrandt established his own workshop to instruct students in his style and technique of painting.  According to guild regulations, the pupils of Rembrandt’s workshop had to paint in their master’s style, under his observation.  (Bruyn 1990, 715).  This supervision may have involved the master providing an underlying sketch over which a student could paint, allowing a student to paint certain features of a commissioned portrait such as the hands, or retouching and finishing a student’s painting. (Liedtke 2004, 57).  The quality and quantity of the work provided by these students under the auspices of Rembrandt have prompted many debates concerning the attribution of certain “Rembrandt” works.  (Wheelock 2014).  Even a Rembrandt signature is not definitive, as Rembrandt may have even signed works painted by an apprentice.  (Liedtke 2004, 53).

Portrait of Rembrandt

The subject of one of these attribution debates is the Portrait of Rembrandt (1650).  In this later portrait, the artist is showing his usual, sidelong glance, but the deeply lined forehead reveals a despondent man beset by personal tragedies.  Unlike most of the self-portraits, Rembrandt in this portrait is wearing an outdated costume: a pleated white shirt, a dark overdress with slashed purple sleeves lined with yellow, and a brown beret worn at a playful angle over a yellow and red skullcap. An inscription at the center right reads “Rembrandt f./1650.”

Art historians have attributed this painting to the workshop of Rembrandt on the basis that certain features point to a pupil painting under the supervision of the master.  The character of the paint mixtures and the types of pigments used are consistent with the practices of Rembrandt’s workshop, but the brushwork lacks Rembrandt’s “sensitivity and vigor.”  (Wheelock).  The hand resting on the staff is inconsistent with Rembrandt’s own modeling of this feature.  Similarly, the costume is executed with thick impastos and broad, flat planes of purple and yellow while the facial features are modeled more delicately.  Generally, Rembrandt did not use two different techniques for the face and clothing in a portrait.  To account for these discrepancies, a workshop assistant likely based the style of the face on Rembrandt’s mid-1630’s portraits and the manner of the clothing on Rembrandt’s style of painting at the time.  (Wheelock).  While Rembrandt likely did not write the inscription, the date and signature indicate that the painting was executed by an artist in the workshop to be sold on the open market.  (Wheelock).


J. Walter Thompson, & ING.  The Next Rembrandt.  (2016). 3D printing on canvas.

J. Walter Thompson, & ING. The Next Rembrandt. (2016). 3D printing on canvas.

The Next Rembrandt

Rembrandt’s style and technique also inspired the second portrait of a man.  Unlike the first portrait, however, this man never existed but is the result of statistical analysis gathered by a computer of Rembrandt’s oeuvre and “painted” using a 3D printer.  This is the painting known as The Next Rembrandt.  Conceived by Bas Korsten of J. Walter Thompson in Amsterdam, in conjunction with ING and Microsoft, this project sought to bring “the great master…back to life to create a new painting” with the use of technology and data.  (ING).  The participants of this project gathered data via 3D scans of 346 Rembrandt paintings, resulting in 160 gigabytes of digitally rendered graphics.  Using this data, they identified demographic elements which occurred in the largest sample of paintings to create a profile of a Caucasian male, with facial hair, of 30-40 years of age, wearing dark clothing with a collar and hat, and facing to the right.  Through statistical analysis they extracted certain features that particularly represented Rembrandt’s style.  Using algorithms to determine the shape and placement of certain features, they were able to identify and replicate dominant patterns that make a Rembrandt look like a Rembrandt.  After finalizing the composition, The Next Rembrandt “artists” created a height map to replicate the brushstrokes and the layers of paint on a canvas surface.  With a 3D printer that outputted multiple layers of paint-based UV ink, the height map regulated how much ink was released on the canvas during the printing process.

Copyright Protection & Authorship

Who is the copyright author of a computer-generated work like The Next Rembrandt?  The Copyright Act extends protection only to original works of authorship fixed in a tangible medium.  Generally, an original work is one “independently created by the author (as opposed to copied from other works)” that “possess[es] at least some minimal degree of creativity.”  (Feist, 499 U.S. at 345).  Copyright vests initially in the author (or authors of the work).  (17 U.S.C. § 201).  While the statute does not define “author,” it is commonly understood to refer to “the party who actually creates the work, that is, the person who translates an idea into a fixed, tangible expression entitled to copyright protection.”  (Reid, 490 U.S. at 737).  

The courts have always grappled with the scope of authorship in the context of “new” technologies.  An 1866 decision, Wood v. Abbott, considered the copyrightability of a work created with the new technology of that time: the camera.  A New York circuit court refused to consider a photograph as a work of authorship, declaring that with photography, “the only force that contributes to the formation of the image is the chemical force of light, operating on a surface made sensitive to its power,” while any human agency was relegated to the positioning of the plate, paper, and frame.  (Wood, 30 F. Cas. at 425).  Almost twenty years later, in 1884, the Supreme Court in Burrow-Giles Lithographic Co. v. Sarony, considered the copyrightability of a photograph of Oscar Wilde.  The defendant argued that there was no identifiable author to entitle the photograph copyright protection because its production was “merely mechanical” involving “the manual operation, by the use of these instruments and preparations, of transferring to the plate the visible representation of some existing object, the accuracy of this representation being its highest merit.”  (Burrow-Giles, 111 U.S. at 59).  The Court, however, held that the photographer was the author regardless of the reliance on a “machine,” which functioned more as a tool to realize the photographer’s creative vision.  In a more recent photography case, the district court in Naruto v. Slater found that the photographer who set up his camera with which a macaque took a “selfie” was not the author of the photograph as he did not actually press the shutter-button.  The court further explained that the macaque who “took” the photograph could not be an author either, as authors have always been only human beings.  (Naruto, 2016 WL 362231 at *3-4).

Outside of the courts, the federal government has addressed the issue of authorship and computers more definitively.  The Copyright Office’s Compendium of U.S. Copyright Office Practices explicitly states that “[t]o qualify as a work of ‘authorship,’ a work must be created by a human being” and that “the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author.” (U.S. Copyright Office 2017, 313.2).  Similarly, the Commission on New Technological Uses of Copyrighted Works (CONTU), established by Congress in 1974, concluded in its report that the computer is not an author but a tool or instrument to facilitate fixation.  More specifically, the Commission stated that:

there is no reasonable basis for considering that a computer in any way contributes authorship to a work produced through its use. The computer, like a camera or a typewriter, is an inert instrument, capable of functioning only when activated either directly or indirectly by a human. When so activated it is capable of doing only what it is directed to do in the way it is directed to perform. (CONTU 1978, 44).

Granting authorship to the computer contradicts centuries of case law and agency determinations as well as the fundamental policies behind copyright.  A computer cannot enforce the exclusive rights granted to a copyright owner or transfer the rights to others through licensing or assignment.  (Denicola 2016, 274).

Who is the Author of The Next Rembrandt?

While the law permits computers as tools in the hands of authors but not as authors themselves, what is the role of the computer that created or helped to create The Next Rembrandt?  Is it acting as a tool, such as a pen, paintbrush, or camera, or is the computer acting independently of any human intervention?  (Bridy 2012, 50).  An art historical framework focusing on the composition and technique may assist with understanding the actual creation and origination of The Next Rembrandt.  The technique behind this painting involved using 3D printing technology to transfer the paint to the canvas like a paint brush or a camera that transfers the image to the paper via a “chemical force of light.”  Thus, this printing technology served more as a tool in the hands of the “artist” to physically realize the concept of The Next Rembrandt painting.  The use of algorithms and data analysis for the composition of the painting presents a more difficult challenge regarding the delineation between human creativity and a computer’s independent actions learned through analyzing data.  The software algorithms may have essentially replaced the artist’s mind and eye in composing this painting, reducing human intervention to insufficient levels for authorship.  Yet, one may argue that the software merely acts as just another tool in the hands of the artist to create and compose this painting.  Taking a step back from the canvas, human intervention occurred in the development of the software as another tool to analyze the data on a selected set of points in order to determine the appearance and placement of a certain set of features.  (Husick 2017).

The presumption inherent in copyright law of a single author behind a work may sway consideration of this problem towards overestimating the involvement of the computer at the expense of recognizing human intervention.  The idea of a single author at the center of copyright law obscures the collective nature of creativity and cultural production, as seen with both The Next Rembrandt and the Portrait of Rembrandt (1650).  (Bridy 2012, 7).  Art history reveals that the artist is not always the one holding the brush or making the composition decisions.  While the Portrait of Rembrandt is signed “Rembrandt,” the master may not have even touched the canvas with his brush.  The students behind The Portrait of Rembrandt analyzed Rembrandt’s work and copied certain features from various paintings to create this particular portrait of Rembrandt.  The programmers/artists behind The Next Rembrandt ultimately worked in a similar way but just with algorithms and analytical “tools.”

Some countries confer authorship of computer-generated work on the programmer.  For example, the UK copyright law grants authorship of a computer-generated work to the “person by whom the arrangements necessary for the creation of the work are undertaken.”  (Copyright, Designs and Patents Act, c. 1, § 9(3) (UK)).  While the U.S. law does not contain such a provision, a broader concept of authorship may be necessary in order to recognize properly the human contributions to works like The Next Rembrandt.  Inspired by the collaborative approach towards the creation of The Portrait of Rembrandt, the concept of joint authorship may be a “legal compromise” or a more practical alternative than bestowing authorship on a single computer program.  A “joint work” is prepared by two or more individuals with the intention that their separate contributions would be merged into a single work.  (17 U.S.C. § 201).  The joint authors of The Next Rembrandt would likely be the software programmers.  In this context, the computer remains a “tool” built by the programmers to create The Next Rembrandt.  Similarly, the concept of a work made for hire, where a person is treated as an author even if they did not play an active role in the actual creation, may also more properly address the authorship issues with The Next Rembrandt.  Generally, a work made for hire is a work prepared by an employee within the scope of his or her employment.  (17 U.S.C. § 101).  The author is the person on whose behalf the work was made, not the actual creator.  The investors or the software companies, who provided the impetus and inspiration for The Next Rembrandt, may more appropriately serve as copyright owners.

The Next Rembrandt Workshop

Copyright law, either intentionally or unintentionally, shapes the creation of art.  Conveying the status of “author” upon an artist allows that artist to control the use of his or her work.  The inherent copyright challenges in The Next Rembrandt, however, reveal a fundamental question about this area of law: should it matter whether an artist uses a paintbrush or an algorithm to create a work of art?  While developers behind The Next Rembrandt did not intend to reduce one of the great artists of the Dutch Golden Age to an algorithm, this technique “offers an opportunity to test your own ideas about his paintings in concrete, visual form.” (ING).  

Perhaps guidance about The Next Rembrandt lies with the actual Rembrandt portraits.  Like the Portrait of Rembrandt, The Next Rembrandt project involved multiple “students” seeking to learn from the style and technique of the great master.  In the twenty-first century, these lessons involved algorithms and data analysis.  Beyond the solitary artist, adopting a wider view of authorship in copyright that instead builds among the concepts of joint authorship and works made for hire may help to reframe the questions of copyrightability for this and other computer-generated works.  A new perspective would allow the technology behind The Next Rembrandt to become part of the toolbox of future artists.


Broos, B. P. J., et al. "Rembrandt van Rijn." Grove Art OnlineOxford Art Online. Oxford University Press, accessed Nov. 6, 2017, http://www.oxfordartonline.com/subscriber/article/grove/art/T071362.

Bruyn, J. 1990. "An Unknown Assistant in Rembrandt's Workshop in the Early 1660s." The Burlington Magazine, Oct.: 714-718.

CONTU. 1978. Final Report on New Technological Uses of Copyrighted Works.

Ekkart, Rudi. 2008. "Portraits in the Golden Age." In Dutch portraits : The Age of Rembrandt and Frans Hals, by Rudi Ekkart and Marieka de Winkel, 17-47. London: National Gallery Company.

Husick, Lawrence A. 2017. Artifical Intelligence and Copyright Ownership. July 7. http://garson-law.com/ask-dr-copyright-artificial-intelligence-and-copyright-ownership/.

ING. n.d. The Next Rembrandt. https://www.nextrembrandt.com.

Liedtke, Walter. 2007. Dutch Paintings in the Metropolitan Museum of Art. Vol. II. New Haven: Yale University Press.

Liedtke, Walter. 2004. "Rembrandt's 'Workshop' Revisited." Oud Holland, 48-73.

U.S. Copyright Office. 2017. Compendium of U.S. Copyright Office Practices. 3.

Wheelock, Jr., Arthur K. n.d. Anonymous Artist, Rembrandt van Rijn/Portrait of Rembrandt/1650. Accessed Oct, 24, 2017. https://purl.org/nga/collection/artobject/1209.

Wheelock, Jr., Arthur K. 2014. "Issues of Attribution in the Rembrandt Workshop." In Dutch Paintings of the Seventeenth Century. Washington: National Gallery of Art.

17 U.S.C. § 101.

17 U.S.C. § 201.

Copyright, Designs and Patents Act of 1988, c. 1, § 9(3) (UK).

Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340 (1991).

Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730 (1989).

Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884).

Naruto v. Slater, No. 15-cv-04324-WHO (N.D. Cal. Jan. 28, 2016).

Wood v. Abbott, 30 F. Cas. 424 (C.C.S.D.N.Y. 1866) (No. 17,938).

Annemarie Bridy, Coding Creativity: Copyright and the Artificially Intelligent Author, 2012 Stan. Tech. L. Rev. 5 (2012).

Robert C. Denicola, Ex Machina: Copyright Protection for Computer-Generated Works, 69 Rutgers U. L. Rev. 251 (2016). 

Brancusi’s Bird in Space: Is it a bird or is it art?

Brancusi’s Bird in Space: Is it a bird or is it art?

Duchamp's Fresh Widow: Copyright Through the Looking Glass

Duchamp's Fresh Widow: Copyright Through the Looking Glass