Submission Title:
The Ghost in the Machine: Why the EU AI Act Cannot Replace the Human Conscience in Translation Ethics

Submission Theme:
Ethics, Justice, and Responsibility in the Age of Automation

Abstract:
Can an algorithm feel the weight of a word? As generative AI and Large Language Models (LLMs) reshape the translation landscape, the industry faces a dangerous illusion: that speed equals accuracy and automation equals objectivity. This paper challenges the techno-deterministic narrative by exploring the hidden ethical crises of the algorithmic turn, positioning the human translator not as a legacy worker, but as the ultimate guardian of justice, cultural nuance, and truth.

While market realities demand hyper-efficiency, the legal and ethical vacuums created by automated translation remain profound. From data sovereignty breaches to the systemic biases embedded in training datasets, algorithms often function as black boxes that homogenize discourse and erase cultural identity. By evaluating the newly established boundaries of the EU AI Act, this study investigates whether top-down regulations can truly bridge the gap between commercial automation and human responsibility. It focuses on high-stakes environments—such as legal, political, and institutional mediation—where an algorithmic error is not just a glitch, but a liability that can compromise human rights and systemic justice.

Moving beyond the binary debate of human versus machine, this presentation argues that the essence of translation is a deeply ethical, political, and empathetic act that AI can simulate but never possess. It introduces a bold framework for "cognitive sovereignty," asserting that human critical reflection and ethical accountability are irreplaceable. Ultimately, this paper serves as a wake-up call for the academic and professional community: safeguarding the human element in translation is not a nostalgic preference, but a critical legal and ethical battleground in the age of automation.