top of page





Header Link One




Artificial intelligence to support human creativity and discovery


Can we use artificial intelligence to support human creativity and discovery? A new trend known as assisted creation has important implications for creativity. On the one hand, collaborative platforms such as those developed as part of the European PRAISE music learning project (Yee-king and D'Inverno 2014) facilitate the acquisition of new creative skills. PRAISE is a social network-based learning platform that involves humans and intelligent software agents providing feedback to a music student on composition, arrangement and performance. Students upload their solutions to a learning unit (on compositions, arrangements or performances) provided by a tutor. Software agents as well as fellow students and tutors then analyze these solutions and provide feedback. In the case of a musical composition, for example, the agent may say: "Your modulation sounds very good, but you could try going up a major third in bars 5 to 8".

During performances, other intelligent software agents compare the student's performances with those previously recorded by the tutor when he or she uploaded the learning unit to the platform. A camera records the student's movements and the software agents also provide feedback on possible incorrect postures. These types of tools, which accelerate the acquisition of skills, are leading to a phenomenon known as the "democratization of creativity".

As early as 1962, Douglas Engelbart (Engelbart 1962) wrote about a "typewriter that would enable the use of a new text-writing process [...] It would allow ideas to be integrated more easily and thus redirect creativity more continuously". Engelbart not only predicted increased individual creativity, he also wanted to increase collective intelligence and group creativity by improving group collaboration and problem-solving skills.

One basic idea is that creativity is a social process that can be enhanced by technology. If we project these ideas into the future, we could imagine a world where creativity is easily accessible and (almost) anyone can write like the great writers, paint like the great masters, compose high quality music and even discover new forms of creative expression. For someone who has no particular creative skills, it is a great relief to be able to acquire them through assisted creative systems. Although this futuristic scenario is still pure fiction, there are already several examples of assisted creativity. One of the most interesting is the assisted percussion system developed by the Georgia Institute of Technology (Bretan and Weinberg 2016). It consists of a robotic arm that enables percussionists to play with three hands. The 61-centimetre-long "intelligent arm" can be attached to the musician's shoulder. It reacts to human gestures and to the music it hears. For example, when the drummer plays the cymbals, the robotic arm plays the cymbals. When the drummer switches to the drums, the mechanical arm switches to the tomtom.

Another result of great interest for assisted creativity is the genre-to-genre transfer of musical style and harmony developed at the SONY Computer Lab in Paris (Martin et al. 2015; Papadopoulos et al. 2016), which helps composers to harmonize a piece of music of one genre according to the style of a completely different genre. For example, to harmonize a jazz standard in the style of Mozart.


Margaret Boden pointed out that even if an artificially intelligent computer were to become as creative as Bach or Einstein, for many it would only be apparent and not truly creative. I fully agree with her on the two reasons she gives for this rejection, namely the lack of intentionality and our reluctance to integrate artificially intelligent agents into our society. The lack of intentionality is a direct consequence of the "Chinese Room" argument (Searle 1980), according to which computer programs can only perform syntactic manipulations of symbols, but are unable to give them semantic content. It is generally recognized that intentionality can be explained by causal relations. However, it is also true that existing computer programs lack the relevant causal links necessary to show intentionality. But perhaps future, possibly anthropomorphic, "embodied" artificial intelligences, i.e. agents equipped not only with advanced software but also with various kinds of complex sensors that allow them to interact with the environment, will contain sufficient causal links to confer meaning and intentionality to symbols.

As for social rejection, this is why we are so reluctant to accept that non-biological agents (or even biological ones, as in the case of Nonja, a twenty-year-old Viennese painter whose abstract paintings were exhibited and admired in art galleries, but whose work was devalued after it was revealed that he was an orangutan in the Vienna Zoo) can be creative, because they have no natural place in our human society and the decision to accept them would have far-reaching social consequences. It is therefore much easier to say that they appear to be intelligent, creative etc. than to say that they are. In a word, it is a moral problem, not a scientific one. A third reason for rejecting the creativity of computer programs is that they are not conscious of their achievements. While it is true that machines have no consciousness and may never think consciously, the absence of consciousness is not a major reason to deny the potential for creativity or even intelligence. After all, computers would not be the first unconscious creators; evolution is the first example, as Stephen J. Gould (1996) brilliantly points out: "If creation requires a visionary creator, how does blind evolution manage to create new things that are as great as we are?


bottom of page