• Home
  • Latest
  • 'Forget Banksy's Shredder. Is Artificial Intelligence the Next Big Thing in Art?'

'Forget Banksy's Shredder. Is Artificial Intelligence the Next Big Thing in Art?'

Posted on 09 November 2018 by Abby Brindley and Emily Dorotheou

'Forget Banksy's Shredder. Is Artificial Intelligence the Next Big Thing in Art?'

As the art world is reeling from the first ever art work to be created live during auction courtesy of Banksy's destructive sense of humour, Christie’s has caused another wave by becoming the first auction house to sell a piece created by an algorithm. "Edmond de Belamy, from La Famille de Belamy" (also known as "Portrait of Edmond Belamy") created by Obvious, a Paris based collective, sold for a whopping $432,500 after having been given an estimate, perhaps rather conservatively considering the novelty, of only $7,000 - $10,000.

This signals the formal arrival of AI generated work and marks a significant step for an industry traditionally perceived as being very conservative. However, whilst technophiles celebrate the art world's apparent embrace of AI generated art, the lawyers amongst us are considering the various legal questions raised by works created in this way.

What is artificial intelligence?

Artificial intelligence (AI) is intelligence demonstrated by machines. It allows machines to exhibit cognitive functions associated with humans such as “learning”, “problem solving” and “decision making” using complex algorithms. Once an AI algorithm has access to a sufficient data set, it will analyse the data, recognise patterns, execute instructions and create an output. When an AI algorithm takes the correct decision or successfully executes an instruction, the device learns from its success. If not, the AI device learns from its mistakes. This process happens multiple times until the AI device starts to make the right decisions or complete instructions autonomously. 

AI and the art world

The AI algorithm that created Portrait of Edmond Belamy consisted of two parts. The first, the "Generator", was given a data set of 15,000 portraits painted between the 14th and 20th centuries, allowing the algorithm to analyse techniques, style, lighting, colouration and perspective, providing it with the information needed to decide how the portrait should look. The second part, the "Discriminator", attempted to identify the differences between human made images and those created by the Generator. The aim was to fool the Discriminator into thinking the Generator's creations were man made. AI enthusiasts will recognise this as an amended version of The Turing Test for establishing artificial intelligence.

This is not the first example of AI devices or computers creating art. In 2016, a group of museums and research institutes, in conjunction with Microsoft, revealed The New Rembrandt, a painting created by a computer in the style of Rembrandt; researchers at the Art and Artificial Intelligence Lab at Rutgers University are working with an algorithm that purports to be creative; a gallery in New Delhi recently held an all AI generated art exhibition; and Turkish artist Memo Akten continues to sell AI generated artworks.

Can AI generated work be considered true "art"?

The question as to whether AI generated work can be considered "art" is complex and hotly debated. Can what humans consider beautiful or thought provoking be expressed as mathematical equations? Can art lovers be inspired by machine generated art? Can AI generated works be expressive?

Effort and novelty give art works scarcity which not only goes to the commercial value but also to our emotional interactions with the piece. Obvious has been quoted as saying that the portraits created by its algorithm are unique and the same result will not be created twice, even if trained on the same data set, but it requires very little human effort to produce. Will these pieces be considered valuable by the market? Will collectors engage with the expression of an algorithm rather than an artist? As the majority of purchases are as a result of passion, aesthetics and a desire to support artists and culture rather than investment, these are questions that will be answered over time by the market.

Who owns the rights in an AI work?

A key issue that has yet to be clarified is who owns the copyright in AI works. The position under English law as set out in the Copyright Designs and Patents Act 1988 (CDPA), is that copyright in an original artistic work is owned by the author (unless assigned in writing). However, the CDPA did not foresee creations of AI. Authorship (and so first ownership) of the copyright in computer generated content is dependent on who undertakes the arrangements necessary for the creation of the work. In the context of AI devices, and assuming the output is "original", it is not clear who undertakes these necessary arrangements and so it is not clear who is the author and first owner of any copyright in the work produced.

Arguably the author, and therefore first owner, of the copyright would be the AI device as the algorithm autonomously created the final work. As algorithms are gaining in sophistication and are able to directly deliver works with minimal human involvement, ownership by the AI device will need to be considered more seriously. While there is currently no legal framework in which to consider the algorithm as the author and owner of legal rights, such rights have been granted to ships and companies, so there is no conceptual issue with bestowing legal rights on an algorithm; in fact, an AI robot was granted citizenship of Saudi Arabia last year.

However, the author (and first owner) could be the person using the AI device. Many artists benefit from the use of machines and teams to produce their final pieces; this does not stop us from referring to the original visionary as the artist, despite the fact they may not be the main executor of that vision. A similar argument could be extended to AI generated works, if we view AI algorithms as simply a means of producing a work requiring human supervision. 

Alternative arguments would be to view the work of the original visionary and the AI algorithm as a collaboration, resulting in joint ownership of the artwork by the visionary and AI algorithm, or to view the AI algorithm programmer as having a right over the final piece if we are reluctant to grant legal rights to the AI device itself.

English law is not alone in having failed to evolve with the times to deal with the problems triggered by artificial intelligence. There are other considerations, in addition to the above, for example whether the output generated by an AI device is even capable of being protected by copyright. The question of copyright ownership is unlikely to present an issue if a human user owns the AI algorithm, but problems may occur if an AI algorithm is owned or licensed by third parties. Parties involved in the creation of AI generated works should be mindful that the law remains unclear and ensure that any agreements relating to machine generated creations include provisions to clarify the rights of all those involved.

How can we help you?

How can we help you?

Subscribe: I'd like to keep in touch

If your enquiry is urgent please call +44 20 3321 7000

Crisis Hotline

Emergency number:

I'm a client

Please enter your first name
Please enter your last name
Please enter your enquiry
Please enter a value

I'm looking for advice

Please enter your first name
Please enter your last name
Please enter your enquiry
Please select a department
Please select a contact method

Something else

Please enter your first name
Please enter your last name
Please enter your enquiry
Please select your contact method of choice