Generative artificial intelligence (GenAI) is being increasingly positioned as a solution to the chronic administrative burden impacting social work practice in England, national workload data evidencing that practitioners are struggling, and are at times, unable to complete their work within contracted hours. This dissertation critically examines how GenAI influences professional decision-making in social work practice under conditions of cognitive overload, employing a hybrid systematic and narrative literature review that brings together cognitive load theory (CLT), social work decision-making theory, and emerging GenAI literature into conversation for the first time.
Using cognitive load theory (Sweller,1988) as the primary analytical lens, this literature review establishes that cognitive overload is a structural, rather than incidental, feature of modern social work practice. It is argued that, while GenAI genuinely offers short-term relief from administrative burden, its positioning as the primary solution to overload, as opposed to a solution, creates the primary conditions for uncritical reliance on its outputs to occur. Considering the empirical evidence surrounding cognitive offloading, as well as the accountability gap between algorithmic output and professional judgement, this dissertation finds that what is actually placed at risk is not secondary to the practice experience, but remains constitutive of it. Intuitive, contextual, values-laden reasoning existing at the core of what defines social work expertise and the development of such. This literature review concludes that social works ethical and governance responses have failed to keep up with the speed at which AI Integration is occurring, identifying a significant gap in practice-based evidence involving social workers directly. As a result, recommendations are made for practice, organisational leadership, and policy which have been oriented around a constructive framework for ethical AI integration, followed by a call for further research to be undertaken on an ongoing basis as the evidence base develops.
Key words: Artificial Intelligence; Generative Artificial Intelligence; Cognitive Load Theory; Cognitive Overload; Professional Judgement; Decision-Making; Social Work Practice; Cognitive Offloading; Accountability; AI Ethics