This presentation argues that contemporary labour-market transformation follows a linked sequence: changes in the form of work, changes in subordination, and changes in employers’ modes of control. Using the concept of "melting labour," it shows how digitalisation, platformisation, and outsourcing dissolve the boundaries that historically stabilised standard employment—between tasks and jobs, working time and non-working time, workplace and home, and employment and self-employment. As these boundaries melt, welfare and labour institutions that remain anchored in standard-employment assumptions struggle to recognise and regulate emerging risks, generating institutional mismatch and welfare state drift in South Korea.

The presentation then traces a shift from visible to invisible control. Whereas industrial subordination was organised through direct supervision, hierarchy, and explicit discipline, platform infrastructures relocate key decisions into opaque, data-driven systems. Through algorithmic control—including allocation, evaluation, task-direction, and penalties such as deactivation—workers can be disciplined without face-to-face supervision and firms can govern labour without issuing explicit commands. This transformation is conceptualised as "invisible control," producing algorithmised dependency that often evades conventional legal and regulatory categories.

Finally, the presentation links these dynamics to the trajectory of South Korea’s youth labour market, where non-standard and platform-mediated entry paths increasingly structure early careers. Rather than treating youth insecurity as a narrow problem of employability or skills, the Korean case highlights how digitally mediated control and fragmented responsibility can reorganise dependency while leaving institutional protections misaligned. The presentation closes by outlining comparative implications for countries undergoing similar transitions: how boundary-eroding work and algorithmic governance interact with dualised labour markets and legacy welfare institutions, and why institutional capacity to recognise and regulate "invisible" forms of subordination has become a central challenge for contemporary social protection and labour-market governance.

How can we explain cases in which effective social policies become neglected by policymakers and citizens alike? Highly effective policies reduce the visibility of the social problems they address (e.g., poverty, poor health, inequality), thereby creating the illusion that the problem has been solved. As a result, beneficiaries may become less politically active, as the urgency that once motivated mobilization declines, and voters may similarly shift their attention toward other issues. These dynamics, in turn, create opportunities for opponents (e.g., austerity advocates, science skeptics) to argue that the policy is no longer necessary or has become too costly. Political support for a social policy may therefore erode because it succeeded, not because it failed. This study investigates this curse of highly effective social policies by developing a theoretical model that captures these self-undermining dynamics and generates several unintuitive predictions about the relationship between a policy’s effectiveness and the temporal evolution of public support and funding.