Reflections on the MIC 2025 Creativity Conference

Following the recently concluded Marconi Institute of Creativity (MIC) conference in Bolzano, Italy only recently concluded, I would like to share some observations and comments with the JCB readership. This year's conference was the 9th organized by the MIC since the initial offering in 2013. It's fair to say that the MIC (or “Mick” as attendees now refer to it) has established itself, arguably, as the premier creativity conference, certainly in Europe, if not globally. Participants represented countries as far afield as Brazil, Canada, the United States, China, and Australia, as well as various European nations. Over 100 creativity researchers attended, sharing their research in three parallel sessions over 3 days. One thing that this conference makes very clear is that there is no substitute for being present. Conferences are not really about the presentations. Those are just an excuse for gathering together. The real magic happens in the spaces between presentations: at morning coffee, at lunch, and in the social events. But these work best when the conference is smaller and more focused. Like many scholars, I've been to big conferences with 1000+ attendees and a bewildering array of sessions, keynotes, and sponsors' booths. These, however, can be some of the least engaging. The people you want to connect with are hard to find and, in the end, the opportunities to talk, exchange ideas, and form potential partnerships are lost. MIC, both this year and in the past, has the balance right. Around 100 people, give or take, means that it is possible, just about, to meet everyone. Three days, and three streams, is enough for some variety, but not too much that people are overly dispersed. Breaks are long enough for meeting people, while a 1-hour lunch means that some good discussions are possible. A social event (with some relaxed keynotes) on three consecutive evenings also engenders a sense of belonging to a small, select group of like-minded researchers, both for graduate students attending their first conference and for veteran professors alike. Finally, another secret is the location. It used to annoy me when conferences were not in a city center location (how can I slip away and do some sightseeing, if I get bored). Now I see it differently. Choose a venue that is just remote enough to encourage participants to stay all day, but not so far that it is difficult to get to. MIC 2025 achieved this balance. Now to the content. While the real work of the conference may happen between presentations, the content of these sets the tone for the discussions, as well as providing a barometer of what is currently hot in creativity research. What was immediately apparent at MIC 2025 was the impact that AI is having on creativity research. Although not the dominant theme in terms of number of presentations, the 16 or so papers dealing with some aspect of AI in creativity show that this topic is important. These papers also ranged across issues such as the automation of the scoring of creativity tests, therefore frequently also touching on education, as well as the question of how best to use AI to support (or augment?) human creativity. I see, however, two reasons for some caution in these focal points. First, while it is true that the ability to conduct fast and accurate evaluations of verbal semantic distance is valuable, and large language models (LLMs) are increasingly being used to automate the Alternate Uses Test, we must not lose sight of the fact that these are only proxies. Verbal semantic distance is a proxy for divergent thinking, and divergent thinking is a proxy for creativity. In other words, we must not forget what we are seeking to measure, no matter how easy AI and LLMs may make these assessments. Most participants at MIC 2025 seem to acknowledge this, and yet, every week we see new papers appearing showing either a minor incremental improvement to an automated DT measure, or a slightly better semantic distance model. There is a serious risk that we will entrench a fundamental misunderstanding of creativity, not among scholars in the creativity research community, but in those outside of the community who look to us to guide their understanding. The second point of caution relates to what already seems to be regarded as a fundamental truth about AI. Let me go back a step for a moment. Three years ago, when LLMs first hit the market, there was the “AI is so creative” movement. That took a couple of years to cool down, as people accepted that LLMs may not be quite as creatively capable as they first thought. It's one thing to be able to spit out uses for a brick (especially when you have probably seen, somewhere in your training, lots of diverse answers). It's quite another thing to produce creative (i.e., novel and effective) outputs to more consequential, unprecedented problems. Unfortunately, the move towards a more balanced view of AI creative capacity seems to have its own untested assumptions. In simple terms, the discussion has now seamlessly shifted to “Human + AI Co-creativity.” The premise here is that AI will definitely make humans more creative. My question is: why? The real research question here should be: will AI, as a support tool, make humans more creative? Maybe it will, but maybe it won't. If, as my own theorizing about the mathematical limits of LLM creativity suggests, LLM creativity peaks at around the human average, then maybe it won't be all that helpful, at least not for everybody. There is some nuance here that needs to inform the research. Put simply, it is this. If you are a Pro-C creative writer, for example, would you really take advice from an LLM that is far less capable than you? Indeed, would you take advice from an amateur human? The same applies in almost any domain of professional practice. If you are a brain surgeon, would you first consult with a high-school biology student before proceeding with your surgery? Probably not. In fact, I hope not! Of course, if you are an amateur in some domain of practice, things look very different. The amateur novelist, struggling for a plot point might, indeed, derive value from an LLM. Not because the LLM is operating at the level of creative genius (and not even because it has the creative ability of a professional writer) but simply because it's just a little better than the human amateur. As the saying goes, “to the blind man, the one-eyed man is king.” Therefore, we have to be very careful not to create the wrong impression from our research. Asking 200 undergraduate students in psychology to write short stories with and without an LLM as an assistant may well show that the LLM as a helper lifts creativity. Indeed, there are already studies showing this. But does that mean that an LLM is a universal creative assistant? No! Such studies mask the fact that we are using amateurs to draw universal conclusions about specific domains populated by individuals with a wide range of abilities. In fact, this issue has probably been lurking in the background of creativity research for a long time. We know a lot about the creativity of certain groups—not least, university students—but not so much about others. If we are going to tackle issues such as “LLM as creative assistant,” we must start to differentiate more between amateurs (or mini/little-c creatives) and professionals. Beyond AI, the MIC 2025 conference showed the greatest focus on creativity in education, with nearly 30 presentations touching on this topic. On top of that, there was also the plenary session on the PISA 2022 creativity data. This valuable exercise, for which the OECD and all the contributing experts must be congratulated, confirms, once and for all, that creativity is important in education. Broadly speaking, there is evidence that countries that prioritize creativity in their curricula do better. However, there is work still to do to raise creativity to the level of attention it deserves around the world. This is not simply because creativity is a nice thing, or even that it is good for self-actualization or well-being. It is important because, as we now see with the rise of AI, the jobs of the future depend on it. Children leaving school in 2025 and beyond face real competition from AI, automation, and robotics. This competition is present across the full spectrum of jobs, from physically routine and predictable tasks (packing items in boxes on a production line) to cognitively routine and predictable tasks (e.g., data processing). The antidote to the challenge presented by AI is creativity. Not only creativity in the sense of being able to focus on higher-order, unpredictable, and non-algorithmic problem solving in existing jobs, but also creativity in the sense of being able to find a creative way to employ yourself in the future. Sadly, even in Australia (which did quite well on the recent PISA creativity tests), there is a reluctance to put rhetoric into action. In Australia, we have a national curriculum authority, and creativity mandated as a general capability. This may have influenced the PISA results in a positive way, but I fear the broader message is not getting through. This is summed up by a conversation I had recently with a high-fee private school in my hometown. After months of discussions with me about creativity and their curriculum, they told me they didn't want to proceed with creativity any further. The reason? They “didn't want to do anything to disturb their current success formula.” In their case, there is no doubt that this statement was a not-very-well-disguised acknowledgement that they see their job as training their students to get into prestigious degree programs such as law and medicine. The school understands that they are a factory for converting parents' money into places in high-status degrees at university. This is as much the fault of parents as it is of the school, especially as they may be preparing these students for jobs that won't exist in the future. Where do we, the community of creativity scholars, fit in? The strong focus on education at MIC 2025 is a good sign, but it needs one further tweak. As I said in a comment in the PISA plenary, I have a prediction to make (intended as a challenge). It is this: In 3 years, we will still be slicing and dicing the 2022 PISA data, rather than acting upon it. In other words, I believe that the discipline must now shift to a stronger end-user focus. Schools, students, and parents will not thank us, in 10 years if all we have done is show that the PISA data confirms that X is correlated to Y. They will not thank us for getting the OECD to do a second round of creativity data collection, in 2026 or 2028. They will, however, thank us for giving them the tools they need to address the challenge posed by AI and the Future of Work. Make no mistake, the work presented at MIC 2025 is a snapshot of a range of interesting, exciting, and potentially transformative scholarship. In general, the discipline of creativity research is fit and healthy! The one weakness, however, is that we are still too inwardly focused. There is an appetite in education systems around the world for creativity, but we cannot afford to take a “technology-push” approach (i.e., we create knowledge and tools that interest us, and the market automatically adopts these). As a discipline, we have to be more focused on responding to market pull. In other words, we need to listen to education departments, schools, teachers, and students and make sure we give them what they say they need to respond to the challenge of AI and the Future of Work. This is our challenge: focusing our research on end-user needs. The author declares no conflicts of interest. Research data are not shared.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.