The Art in Artificial Intelligence

Are machines able to create creative works of art?

The simulation of human intelligence is an important field of computer science. This topic currently triggers many discussions in society. One is about artificial intelligence (AI) and creativity: Are machines able to create creative works of art?

The term artificial intelligence was first used at the Dartmouth Conference in the US state New Hampshire in 1956. It is not easy to find a definition of AI because we don’t even have an exact definition of the word intelligence. In general, we use the term artificial intelligence to describe the field of computer science where human thinking and the simulation of human intelligence via electronic machines is examined. AI basically includes algorithms (computer codes).

The term creativity has a wide range. It describes the inventive potential in acting and thinking, including originality with a link to the solution. Albert Einstein once said: “You can never solve problems with the kind of thinking that created the problems.“ Creativity results in an ambiguity of active doing and passive happening. Basically every human being has creative potential, whether and how much this potential is used depends on various factors.

An interesting perspective whether the art of a machine can be seen as a creative act was offered in the 1990s from Margaret Boden in her book “The Creative Mind”. She championed the view that it is less important if machines are creative or not. The much more interesting question is whether they can create things that humans can recognize as such. Even nowadays we still don’t really understand what human creativity and what creative thinking constitutes of. The assessment of whether a work is creative or not remains a case by case decision. Due to the increasing number of “artistic” works by machines, we should try to find out if our concepts of art and creativity are still up-to-date and whether works of machinery cannot be creative too.

Artificial intelligences already have produced numerous works in the fields of fine arts, literature, poetry, music and games. One interesting example is the portrait “Edmond de Belamy”, created by a computer algorithm. It was sold at the auction house “Christie’s” in New York for about 432,500 $. “Edmond de Belamy” is the brainchild of the French artist collective Obvious, which seeks to democratize art through artificial intelligence. The artist Pierre Fautrel initially fed 15,000 classical portraits into a computer system. From these portraits, the computer automatically generated a series of new images using an algorithm. From them, the collective selected eleven, the so-called Belamy family, which also includes the now auctioned image.

Literature is another area in which AI is currently being tested. Americans have made the attempt to create a fairy tale from Brother Grimm with AI. The result is “The Princess and the Fox”. Computer algorithms that “mimic” the style of classical composers in particular have been around for many years and the results have become more convincing. An artist collective named Botnik has now written the new fairy tale on behalf of a meditation app. A program was fed with the most common words and phrases of the approximately 200 Grimm fairy tales, the artists have composed the fairy tale from the retort. Examples like these were also tested with famous novels like “Harry Potter” or “Game of Thrones”.

To clarify the question of whether art created by AI is worth protecting, copyright must be considered. It says: “author is the creator of the work”. So far, copyright can only be claimed by people. Works exclusively are defined as “personal mental creations”, whereby for the copyright protection a certain “height of creation” is demanded.

Let’s come to the current status of the legal situation with computer programs. Computer programs are protectable; independent products of computer programs, such as paintings and music, are not protectable. The program itself can therefore be protected under § 69a copyright act as a creative achievement of the programmer. Only the work, which generates a computer program independently, is not protectable. Something else would doubtless apply, if a human being would intervene in this process. Accordingly, works created solely by AI do not have any copyright protection, since no human has influenced the final creative process. There are also considerations of protecting algorithmic compositions, such as works by AI and extending copyright to machines. The Legal Affairs Committee of the European Parliament has come up with a draft to classify robots as “electronic persons” and assign them rights and obligations.

In addition to copyright, there are other areas that need to be considered in relation to AI. In order to establish Germany as the world’s leading AI site, there should be a variety of measures to make it easier to make large amounts of data available and usable. Personal rights, the right to informational self-determination or other fundamental rights should not be violated. The general legal framework for algorithmic and AI-based decisions, services and products should be reviewed and, if necessary, adjusted. This is to ensure that effective protection against distortions, discrimination, manipulation or other abusive uses is possible.

The topic of artificial intelligence is also taken very seriously by the government, which is why the new AI strategy for Germany emerged in a comprehensive democratic process. Chances are seen to strengthen research in Germany and Europe, to be drivers of innovation in the world of work and to shape structural changes in the labour market. The use of AI will fundamentally change the day-to-day work of many people. In this change, the Federal Government will support all employees with a variety of measures based on a national training strategy.

The government will promote, analyse, evaluate and critically monitor the use of AI, both at individual businesses and through higher-level observatories, nationally and internationally. It requires a broad understanding of the population about AI, especially among young people, to generate new content and ideas for education, training and further education.

The further development of AI, and in particular the development of strong and superintelligent AI’s, poses opportunities and threats. Machines are already making our everyday life easier in many ways, such as helping to increase our life expectancy. Not every machine needs to be creative to simplify our lives. However there is fast potential in AI created art, even if it can’t be considered as product of a creative process. The potential lies for example in a possible creation and distribution of individual and personalized art and music. In the future AI could make it possible to provide art which considers for example personality and mood of the individual. Furthermore machines can as matter of fact also support artists in their daily work by being a source of inspiration. The Berlin artist Roman Lipski uses an algorithm that acts as his “muse”, so to speak, when he is searching for new ideas. The algorithm picked up his style through pictures of the artist and can now design new pictures based on it. The artist is inspired by the diversity of the AI, according to his own words, the AI helps him to do new paintings and he is now trying out new shapes, colors and styles.

Just as people and machines differ, so do their works of art. Accordingly, we should differentiate in the evaluation of art. Besides the creative works of machinery are, in our opinion, a creative achievement of both – programmer and algorithm. AI research and management must, like any other field, respect the “idea of ​​good and morally right action”. Digitization and robotization can not be considered independently of certain issues. In detail in the development we have to respect ethical, social and ecological principles, values ​​and rules. Even though that means to be patient sometimes and to have a slow progress. The problems that already exist on our planet and the problems that are yet to come require creative thinking. For the future we hope smart people and smart machines will work together in a sustainable and peaceful way and that there will still be enough space for other living things and for human made creativity and art.

A text by Anne Swodenk, Luna Mohr, Rebecca Kaluza, Sarah Licht, Benedikt Mugrauer