Sumitronomics 4.0: Augmentation or Automation?
Sumitronomics 4.0: Augmentation or Automation?
Prof. Sumitro Djojohadikusumo’s thoughts on the importance of human resource quality as a prerequisite for industrialisation now find their most urgent relevance in the era of artificial intelligence (AI). Amid the ceaseless flow of algorithms, his intellectual legacy is not merely an academic note but a stark warning for the future of the national digital economy.
The human capital theory pioneered by Schultz (1961) and Becker (1964) asserts that investment in people through education and training is a productive investment that yields economic returns. However, this principle now faces a heavier test.
Without deliberate paradigm transformation and precise incentive engineering, Indonesia risks repeating the pattern once criticised by Sumitro: a country that grows statistically but lacks structural sovereignty (Djojohadikusumo, 1994).
Two Technological Paths: When the Market Chooses the Easiest Route
The transition to an AI-based economy is not a neutral process but one that is biased. Acemoglu and Restrepo (2019) formalise the fundamental difference between two paths: automation—which replaces labour through a displacement effect—and the reinstatement effect—which creates new tasks so that labour demand recovers.
Simply put, when technology is directed to automate routine tasks, the result is worker displacement and wage stagnation. Conversely, when technology serves to enhance human cognitive and creative abilities—this is what is commonly called augmentation—productivity and skill demand actually increase.
The Future Jobs Report: Robots, Artificial Intelligence, and Digital Platforms in East Asia and Pacific (Arias et al., 2025) notes that in five ASEAN countries, robot adoption between 2018 and 2022 created around 2 million new jobs for skilled workers while displacing about 1.4 million low-skilled workers. Overall, productivity and scale effects outweighed displacement effects, so new technology increased total employment and wages in the region.
However, the same report notes that only about 10% of jobs in East Asia and the Pacific involve tasks that are highly complementary to AI—far lower than in advanced countries, which reach about 30%.
For Indonesia, the gap is estimated to be even wider, given that its workforce structure is still dominated by the informal sector at around 60%. McKinsey (2019) estimates that around 16% of total working hours in Indonesia could potentially be automated by 2030, equivalent to the potential displacement of up to 23 million jobs.
The problem is that Indonesia’s digital economy currently moves more along the automation path due to the skewed market incentive structure: the gap between the cost of capital to purchase subscription AI software and the cost of retraining workers is very wide. As long as there is no intervention to correct this relative price, the market will continue to choose the cheapest solution—shutting down worker positions rather than humanising them.
Autor (2015) reminds us that automation hits hardest at rule-based routine jobs, while jobs requiring creativity, complex judgement, and interpersonal interaction become even more valuable.
The critical question is: which direction is Indonesia’s education system heading—building complementary skills that synergise with AI, or producing workers whose skills will become obsolete in a decade? This is where the state’s role as a market-shaper becomes crucial, as emphasised by Mazzucato (2013). Countries that successfully transform are not those that passively wait for the market, but those that dare to shape it.
Digital Capabilities as a Right, Not Merely a Skill
Sen’s capabilities approach (1999) offers a more humane dimension: technology ideally should expand substantive human freedoms. Digital capabilities are not merely the ability to operate devices, but a fundamental right to access, critique, and create digital spaces without structural dependence on monopolistic platforms.
Nussbaum (2011) asserts that true education must develop reflective capacity and autonomy. In the algorithm era, this means the ability to understand data biases, protect privacy, and participate meaningfully in technology governance.
The education system must move from a front-loaded learning model to an institutionalised lifelong learning ecosystem. The OECD (2025) emphasises that equitable access to high-quality learning from early childhood to adulthood is a prerequisite for inclusive economic growth. Without it, skill gaps will continue to widen.
Arias et al. (2025) add that developing countries need a mix of adaptive vocational training, micro-credentials, and strengthened numerical and digital literacy as a foundation. Equally important is meta-learning ability—the ability to continuously learn how to learn—as the most strategic competency in the disruption era.
Five Pillars of Policy Architecture
Realising sovereign digital humans requires a policy architecture that goes beyond rhetorical discourse. Just as Sumitro was never satisfied with vision without implementation.
First, public investment in STEM infrastructure must be progressively increased and targeted at underdeveloped regions through a National Digital Talent Fund managed transparently with industry partners: allocate the largest resources to areas with the greatest capability deficits, not where infrastructure is already comfortable.
Second, universities and polytechnics must be reformed into living labs directly connected