2024’s Oscars brought emotional contrasts, with blockbusters evoking joy and Oscar nominees leaning toward intense narratives. Using Affectiva Media Analytics and iMotions’ respiration insights, this analysis reveals audience engagement trends, exploring how generative AI influenced award-winning films and shaped emotional responses in this year’s cinematic landscape.
Table of Contents
2024 marked the year of returning popular franchises such as Inside Out, Dune, Despicable Me, Beetlejuice, and Deadpool (and more!) to the silver screen, and the world moved on from seeing Barbie pink everywhere to seeing pink and green of Wicked (because after all, “pink goes good with green”!).
The year also sparked a lot of conversation in the entertainment world as generative AI officially entered the filmmaking space being used for audio editing. And to make things interesting, two films that had used generative AI technologies were nominated for Best Picture and other award categories for this year’s Oscars. Both The Brutalist and Emilia Perez had used genAI – The Brutalist using Respeecher to refine actors’ Hungarian accents and vowel sounds and Emilia Perez to make modifications to a singing voice to hit higher registers.
As awards seasons continue, we are keeping our tradition of reviewing all the entertainment content that we tested over the previous year with our Affectiva Media Analytics video testing solution for entertainment content, comparing across Oscar Best Picture nominees and 2024 top 10 US Box Office hits (see previous analyses done for 2024 and 2023). With Affectiva’s merger with iMotions, we have been able to unlock some more insights into how people are emotionally engaging with content, and see where we can draw comparisons across Academy Award buzzworthy films and those that dominated the US box office in 2024.
Roll Camera! Our Findings And Insights With Facial Expression Analysis for the 2025 Oscar Nominees
When we look at the facial coding data for trailers we have tested in each category, we see that both Oscar Best Picture nominated films and those that were top 10 box office hits performed very well with audiences. Facial expression data from the trailers show that they can maintain high engagement and emotional power, which bodes well for in-market performance once the films are released (below).
Dissecting emotional engagement further, we also looked at the data for strong positive reactions (think: Smile, Joy, Sentimentality) to discern whether audiences were positively engaged with the trailers. We saw that US Box Office blockbusters are nearly four times more likely to evoke strong positive emotional reaction compared to Oscar Best Picture nominees. This finding remains consistent throughout the years, with minimal changes (below).
Keeping in line with similar trends from previous years, we found that box office blockbusters showed higher levels of Joy and Sentimentality, and those that were Oscar nominees were more likely to have more negative responses (think: indicators of Sadness such as Brow Furrow, Downturned Mouth). Below shows our correspondence analysis, mapping out different emotional state responses and the association with each category.
One reason for this may be attributed to genre and what is more likely to frequently show up on box office hits or Oscar nominee list – on average, we see that top box office top ten lists tend to have a larger presence of family-friendly comedy or animated films good for children (e.g., Despicable Me 4, Moana 2), as well as action, fantasy or superhero movies that have a strong fan following (e.g., Deadpool & Wolverine, Antman and the Wasp in recent years), that all play a part in generating high positive engagement.
In comparison, films that are Oscars nominees tend to take more of a darker tone, eliciting more sadness and negative reactions (though remember, negative isn’t always a bad thing depending on the context!) to build on drama, suspense and storyline complexity.
What Happens If We Add Physiological Respiration To The Mix?
Adding a new dimension to the analysis, we used iMotions’ Webcam Respiration on respondent face videos to detect subtle body movements associated with breathing. Respiration can be used to help detect physiological arousal – a beneficial addition to our toolkit to investigate whether audiences are emotionally connecting with media.
For this analysis, we took a subset of trailers from each of these categories to preliminarily explore if there were any comparisons to be drawn between Oscars Best Picture nominees and box office blockbusters. When we look at our facial coding Valence measure (net positivity from the face) and average respiration rates (breathing), we see that while blockbusters seem to evoke a variety of emotional states, as indicated by an even spread of cases with both high and low valence and arousal.
However, Oscar nominees tend to have a different pattern, skewing more towards feel-good or intense categories. As we saw in previous analyses, this is likely due to the different narratives that tend to attract Oscar nominations, with more serious or charming movies often getting the nod, vs the more spectacular blockbusters.
We can also see here the benefits of a multimodal analysis of viewer response. When we add respiration into the analysis, it adds a level of nuance that may help us better understand the Oscars trailers in particular. While perhaps the trailers for the Oscar nominees may not have been as emotionally positive, we still are able to see with the respiration data that even with low positivity, there is still emotional intensity to these trailers – something that may not have been as apparent before with just the facial coding data.
And That’s A Wrap!
As we continue testing more entertainment content, it leads us to have more interesting insights into the shifts in moviemaking and types of films that are released, especially when it comes to films that receive award nominations. In 2023 when we began this journey, we commented that Everything Everywhere All at Once was the catalyst of creating a genre of its own, combining humor, drama and action with all its suspense and whimsical elements.
2024 marked the year of Barbenheimer with their two very distinct different genres going head-to-head with each other, and Oppenheimer securing the Best Picture award. And now in 2025, most of the Academy Awards best picture nominees leaned more towards serious and darker themes (the exception perhaps being Wicked), with independent film Anora taking the Best Picture award home.
Regarding trends to watch out for, we may see films using generative AI continuing to pop up in the top 10 and award nominee lists, especially as genAI solutions improve over time. The question will be how filmmakers incorporate this technology into their craft, whether they are transparent about their genAI usage, and if the content retains a strong emotional connection to the viewer in its visuals and storytelling.
What will be the top films for 2025? Who will win the Academy Award for Best Picture in 2026? Make sure to stay tuned for next year’s analysis!
Behind the Scenes: Analysis Methods
We analyzed a selection of the trailers for Oscars Best Picture nominees between 2022-2025, corresponding to films that were released between 2021-2024. For the blockbusters, US Domestic Box Office top performers in 2021-2024 were also analyzed. We then performed a deep search in Affectiva’s database for trailers and teasers of these films that were tested with our facial coding technology in the US market, pulling out summary metrics and normative data for these movies. Respiration data was obtained retroactively by reprocessing face video sessions for a selection of projects, namely those that had scored high in expressiveness.
Average duration of each piece of content ranged from 30 seconds to 3 minutes in length. Note that films that were both Box Office top performers and Oscars Best Picture nominees were counted in each of their categories, with approximately 45 films being included for the analysis. For each film, emotional norms for every facial expression and emotional state were averaged across each category list, comparing to each other and to our norm for all trailers tested in the US market. For respiration data, projects’ average respiration rates were averaged across categories.
