Normativity and Productivism: Ableist Intelligence? A Degrowth Analysis of AI Sign Language Translation Tools for Deaf People

Problem
This paper addresses the gap in the literature regarding the ethical implications and societal impacts of AI sign language translation tools, particularly their inherent biases and lack of representation from the Deaf community. It critiques the prevalent use of these technologies, which are often developed without input from Deaf users, leading to systems that fail to accurately capture the nuances of sign languages. The authors argue that existing AI models perpetuate ableism by standardizing communication in a way that marginalizes Deaf culture and experiences. This work is presented as a preprint and has not undergone peer review.

Method
The authors employ a qualitative analysis framework, drawing on philosophical concepts from Jacques Ellul, particularly “The Technological System” and “Technological Bluff.” They critique the methodologies used in developing AI sign language translation tools, emphasizing the reliance on biased datasets that do not reflect the diversity of sign languages. The paper does not present a novel architecture or empirical results but instead focuses on the theoretical implications of current practices in AI development. The authors argue that the existing models prioritize productivity and efficiency over authentic communication, leading to a misrepresentation of Deaf individuals’ lived experiences.

Results
As this paper is primarily theoretical and does not include empirical experiments or quantitative results, it does not provide specific performance metrics or comparisons against established baselines. Instead, it highlights the qualitative shortcomings of current AI systems in capturing the richness of sign languages and the cultural context of Deaf communities. The authors emphasize that the existing tools fail to meet the communicative needs of Deaf users, which they argue is a significant oversight in the design and implementation of these technologies.

Limitations
The authors acknowledge that their analysis is limited by its qualitative nature and the lack of empirical data to support their claims. They do not provide specific case studies or examples of AI sign language tools, which could have strengthened their argument. Additionally, the paper does not explore potential solutions or alternative approaches to developing more inclusive AI systems. An obvious limitation is the absence of a detailed examination of specific AI models or datasets that exemplify the issues discussed.

Why it matters
This paper has significant implications for the development of AI technologies aimed at enhancing communication for marginalized communities. By framing AI sign language translation tools as “Ableist Intelligence,” the authors call for a reevaluation of how these systems are designed and implemented. The critique encourages researchers and engineers to consider the ethical dimensions of AI development, particularly the importance of including diverse voices in the design process. This work advocates for a shift towards more inclusive practices that prioritize the needs and experiences of Deaf individuals, ultimately aiming to foster better communication and understanding between hearing and Deaf communities.

Authors: Nina Seron-Abouelfadil, Poppy Fynes
Source: arXiv:2604.28125
URL: https://arxiv.org/abs/2604.28125v1