Exploring professional translators’ attitudes towards control and autonomy in the human-centred AI era: quantitative results from a survey study

This paper presents a survey study on attitudes of US professional translators towards self-perceived “control” and “autonomy” regarding translation technologies. In the human-centred AI (HCAI) and “intelligence augmentation” paradigms, humans retain the highest possible levels of control and autonomy alongside high levels of automation. The rationale for this study is that the cycle of translation technology adoption normally implies human adaptation to technologies that have been previously developed without their input. This often leads to resistance to adoption and negative or divergent attitudes. Now that AI app integration is still at the early stages of development and adoption, it is of utmost importance to identify users’ needs and attitudes to develop tools that professionals can easily adopt and feel in control of. Methodologically, the study involved a self-administered online Qualtrics survey that was completed by 41 US-based translators in May 2024. The self-reported levels of control and autonomy are generally high, while subjects reported medium levels of forced technology use by external agents. The use of generative AI remains low, with 0% of respondents indicating being forced to use AI or LLMs, in line with other recent studies. Future expectations of control in the AI era declined dramatically, but this perceived loss of control in the AI era was attributed to human agents in the process rather than AI apps or algorithms (big tech, developers, LSPs, project managers, clients, etc.). These low levels of sense of agency in the future correspond with recent studies that find that exploitation through “digital Taylorism” by other human agents might be perceived as more of a threat than AI or translation technologies themselves.