Exploring the Use of Artificial Intelligence in Archival Activities

<p class="abstract" style="margin-bottom: 12.0pt; line-height: normal;">This study reports findings from a survey of working archivists examining their use of generative AI, along with perceived benefits, concerns, and barriers to adoption. While respondents expressed general skepticism toward AI, nearly half reported using it at least monthly. AI was most commonly viewed as useful for tasks such as metadata creation and transcription, while its application to appraisal was largely rejected and opinions on reference services were divided. Key concerns centered on algorithmic bias and the ethical implications of delegating decision-making to automated systems, whereas copyright issues were considered less important. The primary barrier to adoption was identified as the lack of accuracy and reliability in AI outputs. Despite these concerns, respondents indicated a willingness to engage with AI if greater transparency were provided regarding software design, training data, justification for results, and if human oversight remained central to its use.