This year, my work took me to the wonderful city of Rio de Janeiro for the 18th edition of the International Conference on Arts and Cultural Management. More than 260 arts management researchers from 31 countries gathered to exchange best practices in the field and to share the results of their research projects.

Personally, I had the chance to present, together with my colleague Caroline Marcoux-Gendron, the results of our study on the attendance of arts activities by audiences from immigrant backgrounds, as well as to attend Vivianne Magnan-St-Onge’s very first academic presentation. Vivianne is a Ph.D. student at ESG UQAM and a graduate of our master’s program in marketing.

What’s fascinating about this conference is the sheer variety of topics covered. In a single day, I got to learn about the storytelling mechanics behind the Taylor Swift empire (during a presentation by Paola Signori), discover — thanks to doctoral student Irene Tsitse — that the challenges of integrating business intelligence into the cultural sector in the Netherlands are essentially the same as those in Quebec, and see just how closely the video game and artificial intelligence industries have been intertwined — and that from their very beginnings³As the Veritasium team explains in this video, we ultimately owe the solution to the protein-structure-prediction problem to the tools DeepMind developed to win at Go, chess and StarCraft II. (thanks, Pierre-Jean Benghozi).

Much as I did in 2024 for the AIMAC conference in Lisbon, I’m sharing here a few highlights from the presentations that stood out to me.

As part of her partnership research project — carried out jointly with CAPACOA, Orchestras Canada, PACT and the National Arts Centre — Élodie Santos looks at the key role that certain consumers can play in the cultural sector: that of initiators.

Initiators are individuals (not affiliated with any cultural organization) who decide, for a variety of reasons, to organize cultural outings with people from their circle or community. Sometimes the initiator is an uncle who regularly organizes theatre outings with his nieces; sometimes it’s a resident of a seniors’ residence who organically plans (without any direct support from the residence) monthly museum visits with her neighbours.

According to Élodie Santos, initiators represent a major opportunity for cultural organizations looking to build audience loyalty, thanks to their closeness to — and influence within — their micro-communities.

Drawing on the results of the first qualitative phase of her project, Élodie Santos proposed a typology distinguishing four main types of initiators based on their motivations (intrinsic or extrinsic) and on how much importance they place on either the artistic performance itself or the social experience surrounding it. For some, an outing is a success first and foremost if the show is of high quality; for others, it’s the social dimension of the activity that matters most. Overall, I found this a really interesting project, because it foregrounds another kind of influence (an initiator is neither an opinion leader nor an influencer!) — one that is well within reach of arts organizations.

In this compelling presentation, the team of authors set out to answer the following question: how do you make climate-change fiction films that are both socially relevant AND commercially viable?

To answer their question, the team identified 60 films belonging to the Cli-fi genre (if you’re as much of a newcomer to this as I was, the term Cli-fi refers to films that touch — closely or loosely — on climate change, and it covers works like Erin Brockovich and Don’t Look Up as well as blockbusters like Avatar and Interstellar). Yazgulieva and her colleagues then used NLP⁴Natural Language Processing. tools to characterize the various persuasive elements found in the films’ screenplays, audio descriptions and synopses.

Overall, I took away two key findings from this study. First, for a film to be both a commercial success AND a social success (i.e. a film that sparks a lot of discussion), it has above all to be entertaining. In short, for a work to have an impact on society, it first needs to be intrinsically interesting. Second, productions that involve scientists tend to be more successful — both at the box office and in terms of discussion — whereas activists only have an impact on social engagement (and not on a work’s commercial performance). Do scientists come across as more objective to the general public? To be continued!

For this third presentation summary, we now turn to the integration of artificial intelligence into advertising agencies, drawing on a major study led by Silvia Cacho-Elizondo and her colleagues in partnership with IAB Mexico (which is, in a way, the Mexican equivalent of our own Quebec A2C).

For this study, the research team surveyed 97 senior managers at Mexican advertising firms about their perceptions of both the benefits and the barriers of using AI in their agencies. One particularly interesting aspect of this work is that it offers a perspective on AI other than the usual Anglo-Saxon one (which is probably heavily biased toward Anglo-American knowledge) — something that naturally resonates with my own francophone, Quebec point of view.

Overall, the study’s results suggest that agencies want to (or already do) actively integrate the various artificial-intelligence tools into their practices, but the benefits of doing so aren’t yet entirely clear. Of course, AI lets many agencies produce content faster and better organize their work (emails, contracts, etc.), but a number of managers are seeing a differentiation problem emerge: if everyone uses the same tools (mainly ChatGPT), there’s a strong risk that campaigns will all start to look alike.

What’s more, according to Cacho-Elizondo, several significant barriers stand in the way of full, lasting AI adoption in the Mexican advertising sector — whether a shortage of qualified staff, insufficient workforce upskilling, or a lack of budget (an issue likely to grow with the new token-billing methods). The solution? Develop new AI maturity models (analogous to digital maturity models), share best practices more widely, and improve training programs. On that note, ESG UQAM recently launched a graduate diploma (DESS) in digital marketing and AI integrated into marketing.

Since the 2000s, the featuring phenomenon (one artist’s contribution to another’s work, credited with the term feat.) has exploded in the music industry; in the global top 100, we’ve gone from a featuring rate of about 5% to 30% today.

Why are collaborations so common these days? We can obviously guess that many musicians are driven by artistic considerations, but we also know that features can have a significant cross-promotional effect on songs (as McKenzie et al. showed in a 2021 article in the Journal of Cultural Economics): for instance, a collaboration naturally increases the number of « pages » a song appears on across Spotify, which ultimately helps it reach a wider audience.

Now, while some authors have already looked into how effective musical collaborations are, we still don’t have a good handle on which collaborations are the most likely to happen in the first place. That’s precisely the question Montoro-Pons, Palma-Martos and Cuadrado-García set out to study. The team of Spanish researchers assembled a large dataset of nearly 8,000 songs produced by a little under 2,900 artists, in order to identify the characteristics that make collaborations most likely.

One of their main questions was whether collaborations are driven first and foremost by a simple network effect (artists choosing their collaborators mainly because they already know them, closely or loosely) or whether it’s artistic compatibility — in other words, similarity — that matters most.

One aspect I especially appreciated in Montoro-Pons and his colleagues’ presentation was precisely the data used to characterize the distance between two genres (a genuinely tricky challenge I’ve run into myself in the past): namely, the tags left by music lovers on the Last.fm platform.

The main takeaway from their study: collaborations are driven first and foremost by aesthetic and linguistic proximity. In other words, an artist making French-language trap-rap will naturally tend to collaborate with other francophone trap artists. Personally, I find this result particularly ironic, given that we now know — thanks to an article by Ordanini, Nunes and Nanni (2018) — that it’s precisely the collaborations between the most aesthetically distant artists that tend to be the most successful. In short, artists would do well to venture off the beaten path more often.