DXspark participates in international study on AI’s impact on audiovisual content interpretation

DXspark conducted the statistical analysis for a groundbreaking study developed as part of a PhD research project at the NOVA Institute of Communication (ICNOVA).

The research reveals how Artificial Intelligence (IA) systems simplify and standardize the interpretation of human-made works, creating what researchers call "meta-identities."

The study is the result of a high-level academic and technical collaboration, featuring researchers from the following institutions:

• • NOVA Institute of Communication (ICNOVA)
  • University of São Paulo (USP)
  • NOVA Institute of Philosophy (IFILNOVA)
  • Faculty of Communication, University of Castilla-La Mancha (UCLM)
  • DXspark

We actively provided technical and analytical support for this mixed-methods research, which compared how humans and AI systems (such as ChatGPT and Gemini) classify and interpret audiovisual content. The project involved an international consortium of experts from Portugal, Brazil, and Spain.

Henrique Carvalho, Data Scientist at DXspark, carried out the detailed statistical analysis that underpins the study’s findings. By applying weighting and normalization models, we ensured comparability between human evaluator data and algorithmic outputs.

This analysis was pivotal in demonstrating a consistent structural divergence between human sensitivity and the procedural logic of machines.

The full article, titled "Meta-Identity and Algorithmic Mediation on Digital Platforms: A Comparative Analysis of AI–Human Content Categorization," is available for reading here.